Strumia’s lament

October 6, 2018

On September 28, 2018, Professor Alessandro Strumia gave a talk at CERN at a conference on “High Energy Physics and Gender.” In the talk, he took the unpopular view that the reason that there are fewer women than men in physics is primarily because of inherent differences between men and women (on average) as far as talent and interests are concerned, and not because physics is oppressive of women.

The organizers promptly distanced themselves from his talk, and removed the video recording and his slides from public access. CERN, where he is a research leader, suspended him, and his home university of Pisa put him under investigation. The ERC is considering to retract Strumia’s 1.9 million euro grant. While his talk is not available for viewing, his slides can be found online. The media have reported a general outcry and wide condemnation of the talk. A collection of physicists published an open letter, signed now by over 1600 people, in which they express their anger, and state that Strumia’s ideas are unsound.

I should mention that the impression that the media give that the whole of the scientific world condemns Strumia is incorrect. I examined some of the twitter feeds which followed the talk, and I found that besides a minority which vocally condemns Strumia and another minority which vocally supports him, the reasonable middle states that Strumia presented facts, and that they would like to see counterarguments against either the facts or Strumia’s interpretation of them, before dismissing Strumia’s statements.

What I find striking about all of this, is that most people who condemn Strumia have not witnessed his talk. For instance, the open letter starts with “[t]he statement here is based upon widely reported events, publicly available slides, and eyewitness accounts.” I.e., the only input that the writers used which is not hearsay are Strumia’s slides. I find that a weak basis for publicly raking someone over the coals.

It has to be said that Strumia’s slides give the impression that he has an axe to grind, and that rather than limiting himself to objective facts, he spent a considerable portion of his talk on his personal experiences, on politicizing his ideas, and on insulting his audience which consisted for the majority of young female physicists.

Within the slides, however, there is also some solid research. Basically, what Strumia does is examining scientific quality based on an objective measure, namely number of citations. Using a very large dataset, he makes comparisons between men and women in physics in relation to citations, and shows that with regards to citations on average women underperform compared to men. You may rightfully argue that citations do not give a complete picture of scientific quality. However, there are few other objective measures which you can use for such research, so at least he provides a factual starting point for a discussion. The question is: how do you explain the observations which Strumia makes?

Strumia compares the “mainstream” explanation (“all the differences between men and women are culturally determined and thus physics is oppressive of women”) with the, what he calls, “conservative” explanation (“women are inherently less interested in physics, and the people with the most talent for physics are predominantly men”). Neither of these explanations can be shown to be “the correct one,” but Strumia at least gives some indications on why the mainstream explanation can be considered faulty. The most damning argument against the mainstream explanation is the “gender equality paradox,” which entails that the more a country does to erase the cultural differences between men and women and the more it does to erase the barriers that women face to make free choices in their careers, the fewer women choose a career in science and technology. This observable fact falsifies the notion that there are no differences between the interests of men and women.

Strumia also rightfully wonders why people are so concerned about the fact that men form the majority of the people in the STEM fields, while nobody cares that women form the vast majority in education, psychology, the humanities, and medicine. He also notes that “equal representation of women” in a field that is dominated by men is demanded where it concerns STEM, but as soon as it concerns jobs that are dirty or dangerous (such as construction, firefighting, or mining), the fact that almost no women are found in these jobs is not seen as a problem at all.

Unfortunately, these sensible statements are overshadowed by Strumia’s wailing about his personal experiences, his suggestion that physics is a men’s job, his complaints about the way institutes tend to assume that men have no issues and that all women are oppressed everywhere, his annoyance with the widespread support that women get but men lack, and his explicit mocking of domains such as gender studies. Mind you, there are grains of truth in his wailing, but he should not have included it in his talk as it undermines the other things he says.

Strumia should simply have presented observable facts, and then either leave the interpretation up for debate, or weigh his own arguments against the arguments of “the other side.” While simply ignoring counterarguments is something that a majority position can do without penalties, Strumia is talking from an underdog position (especially at this conference), and as such he has to be much more careful about how he explains his ideas.

The open letter I referred to does provide some interesting counterarguments to Strumia’s statements. Some of these are convincing (in particular where they concern Strumia’s obsession with citations), others much less so (for instance where they refer to “unconscious biases” and where they attempt to dismiss the “gender equality paradox” by stating that oppressive countries have more women in STEM because of a lack of choice, which only underlines Strumia’s statements).

But is does not matter whether your sympathies lie with Strumia or the open letter (or neither): the point is that both Strumia’s statements and the statements of the open letter should be up for debate. And that debate is not going to happen because (a) CERN tried to erase all of Strumia’s materials, making it impossible to know what he actually said, and (b) Strumia immediately got punished so harshly for expressing his ideas that few people will be willing to examine his statements objectively, as they know that they face similar punishment if they publicly reach the same, “abject” conclusion.

The open letter, in the first paragraphs, states in bold font: “We write here first to state, in the strongest possible terms, that the humanity of any person, regardless of ascribed identities such as race, ethnicity, gender identity, religion, disability, gender presentation, or sexual identity is not up for debate.” I find this text rather unworthy of serious scientists.

The statement makes clear that the authors of the text want the reader to believe that Strumia was attacking the humanity of women in physics. As far as his slides are concerned, there is absolutely no evidence of that, unless you assume that stating that there are inherent differences between men and women amounts to attacking someone’s humanity. You will find that most people believe that there actually are inherent differences between men and women, and rightly so, as these differences can be observed. Claiming that such differences do not exist is dogma and not science.

Now, the statement gives the strong impression that it considers a number of topics as “not up for debate” at all. Because as soon as you bring one of them up, you are probably going to be condemned for attacking someone’s “humanity,” and such topics may not be debated. To declare a topic so sacred that it cannot be debated is as unscientific as you can get. You may find a person’s statements despicable, but if they are supported by falsifiable claims, then your responsibility as a scientist is not to dismiss them outright (and penalize the perpetrator), but to have the debate, offer counterarguments, and if possible falsify the claims.

Unfortunately, Strumia is a proponent of reason and free speech that we could have done without. He is a man who let his frustration with certain trends in science and society get the better of him. The best thing you can hope for with a talk like this is that at least certain topics are opened up for debate rather than being untouchable. However, the steps his opponents took, making him suffer personally for daring to speak his mind, ensured that fewer people will be willing to speak up about these topics in the future.

Advertisements

Identity politics and retirement

August 1, 2018

Researchers of the Dutch demographical institute NIDI argue in an article that retirement age should be lower for people who have a relatively low level of education than for people who have a high level of education. They would like to introduce three categories of people, based on their level of education. In their proposal (depending on the chosen scenario) the category of people with the lowest level of education retires 4 to 5 years earlier than the category with the highest level of education, and the middle category will end up somewhere in the middle. They state that there are two reasons for this proposal: people with a low level of education tend to start earlier with working, and they tend to die earlier.

In summary, they argue for applying identity politics to the setting of the age of retirement; i.e., they divide people into groups along some demographic criteria (in this case education level), make a comparison in some areas between the averages of the groups (in this case life span and age of entering the job market), and distribute advantages and disadvantages over people based on the group in which they end up (in this case with respect to retirement age), regardless how distant each individual person is to the average of the group they are assigned to. That this approach is an aberration is clear from the fact that according to the proposed system, an academically educated sociologist who cannot get a job in sociology and thus works in construction, retires at a far later age than an uneducated person who works in construction.

The researchers make at least two basic mistakes. The first mistake is that “correlation does not imply causation” (as this is a mantra for social scientists it surprises me highly that they make this mistake). The fact that someone has a low level of education is not the cause for them dying earlier. The reason that people with a low level of education tend to die earlier is that they tend to live more unhealthy lives — on average, they smoke more often, drink more, use drugs more often, eat more unhealthy food, etcetera. Also, some of the jobs that low-educated people hold are quite detrimental to physical health (e.g., jobs in construction). But that does not mean that everybody with a low level of education lives unhealthily or works in a physically demanding job. Neither is the starting age of work life necessarily lower for a low-educated person than for a high-educated person (in fact, unemployment is higher among low-educated people than among high-educated people). Moreover, despite the fact that on average high-educated people indeed start their work life later than low-educated people, that does not entail that they work less hours in total in their lives (in fact, high-educated people tend to spend much more time at their jobs than low-educated people).

The second mistake is that the level of education is not the demographic factor that has the biggest impact on how long people live. The factor that has the biggest impact is sex. Women live, on average, 7 to 8 years longer than men. If you are going to apply identity politics to retirement age, the logical first division that you have to make is splitting men from women, and then add 7 years to the retirement age for women. I am sure that the researchers from NIDI know this. However, I fully understand why this fact never comes up in the proposal of the researchers: bringing up this fact would be seen as “sexism.” But realistically speaking, if you want to use demographic factors to make retirement age more “fair” with respect to distribution between years of work and years of retirement, women should retire much, much later than men.

I can see that there is a “problem” in setting the age of retirement purely on the basis of birth year, in that the system is, from some perspectives, not fair. But the solution is not to apply identity politics on a grand scale, on the basis of level of education. As retirement basically should account for the fact that at some point people are too old to work in a particular job, the content of the job should be the only factor in determining retirement age. Therefore, a solution is to make age of retirement part of work packages. A possible implementation for such a solution is as follows:

The “standard” number of work years could be set to 40, or 480 months. You can retire when you have worked that long. However, in certain jobs, 12 months of work would be counted as longer, while in others it would be shorter. Any year after 20 years of age in which one does not work is counted as 9 months. If, for instance, a year of work in construction is counted as 14 months, someone who starts working in construction at 20 years of age, would retire at 54. If a job in academics would count a year of work as 10 months, someone who starts in academics at 25 would retire around 69. Without a job at any time in one’s life, “retirement” would begin at 73. Such a system would account for the kind of job that people do, the number of years that they work in different jobs, the age that they start working, the number of years that they work, etcetera. The most important advantage is that it would relate retirement age purely to the content of the work that someone does.

NIDI examines correlations between certain demographics and certain social facts. That is fine. But when they then notice a correlation which might make one think “it looks like some demographic is getting it in the shorts,” the solution is not to apply identity politics, create some demographic groups, and reward certain groups and punish others. The solution is to determine the cause of the “injustice” and try to deal with that cause. Or, if the cause proves to be an individual choice rather than a systemic issue, shrug and tell people that certain individual choices have negative consequences — if you want to be pedantic about it.


Terechte klachten

May 5, 2018

(This post is in Dutch as it relates to several articles in Dutch.)

Aleid Truijens is schrijver, recensent, biograaf, en voornamelijk bekend als columnist bij de Volkskrant. Ze schrijft over diverse onderwerpen, waarvan “educatie” er een is. Ik lees haar columns altijd met interesse. Meestal ben ik het eens met de strekking van haar verhaal. Zo ook met het stukje dat ze schreef in de Volkskrant van 27 april 2018, getiteld “Ik hoop dat de universiteit een intellectuele vrijplaats is, maar ik ben er niet zeker van.” De conclusie die ze met haar betoog bereikt, onderstreep ik volledig. Maar ze maakt een aantal bochten waar ik grote vraagtekens bij zet, en waarvan het me verbaast dat ze kennelijk de betreffende denkwijze omarmt. Bij deze mijn commentaar.

Aleid’s aanleiding voor haar betoog is een hetze van een aantal studenten van de Universiteit van Amsterdam tegen hun nieuwe “diversity officer,” Anne de Graaf, gepubliceerd in Het Parool als antwoord op een interview met De Graaf in Trouw. De Graaf, die veel ervaring heeft met het thema “diversiteit” binnen Amerikaanse universiteiten, is onder andere aangesteld naar aanleiding van een fel rapport over de “diversiteit” binnen de UvA. Dit rapport, opgesteld onder leiding van activiste Gloria Wekker, schuift “witte mannen” allerlei vermeende misstanden binnen de universiteit in de schoenen, en pleit voor quota. De studenten eisen op hoge toon dat De Graaf afstand neemt van bepaalde uitspraken, zoals haar weerzin tegen quota. Ze is ten slotte aangesteld om inhoud te geven aan de bevindingen van het rapport, nietwaar? Dus waarom stelt ze zich niet net zo op als de activisten die het rapport in eerste instantie hebben geschreven?

Kennelijk hadden de studenten verwacht dat ze met hun “diversity officer” eindelijk de betreurenswaardige slachtoffers van het witte, mannelijke schrikbewind aan de macht zouden kunnen brengen. Dat De Graaf een objectief en realistisch beeld van de werkelijkheid heeft, stoort hen in hoge mate. De studenten meten “rechtvaardigheid” af aan “percentages gemarginaliseerde groepen,” terwijl De Graaf expertise voorop wil stellen. En hoewel de studenten zeggen dat “quota een laatste redmiddel zijn,” stellen ze ook fijntjes dat “er voldoende bewijs is dat quota werken” (waarbij ik denk: natuurlijk, als je enige doel is dat er meer medewerkers van een bepaalde demografie zijn, en je niet geïnteresseerd bent in andere gevolgen van je maatregelen, dan is het instellen van quota ongetwijfeld effectief).

Aleid Truijens begint met een kort historisch overzicht. “Drie jaar geleden stelden de Maagdenhuisbezetters terecht dat de universiteit te wit en te mannelijk (de docenten) was, en het curriculum ‘te westers’. Wat er wordt gedoceerd is voornamelijk wat witte mannen eeuwenlang hebben bedacht.

Daar gaat mijn eerste alarmbel af. Wat bedoelt Aleid met dat woordje “terecht?” Is zij het eens met de Maagdenhuisbezetters? Je kunt stellen dat het op zijn minst opvallend is dat de wetenschappelijk staf gedomineerd wordt door mannen, terwijl de studentenpopulatie gedomineerd wordt door vrouwen, maar dat is onder andere een gevolg van het verleden waar veel minder vrouwen studeerden en van het feit dat mannen gemiddeld (met nadruk: gemiddeld) ambitieuzer zijn in hun werk dan vrouwen. Je kunt eventueel beargumenteren dat het belangrijk is dat er een redelijk percentage vrouwen in de staf vertegenwoordigd is, in ieder geval als rolmodel voor de ambitieuze vrouwelijke studenten die een toekomst in de wetenschap willen. Ikzelf denk dat dat percentage er al is, maar je kunt er een discussie over voeren. Dan blijft over “te wit” en “te westers.”

Hoezo “te wit?” De overgrote meerderheid van de Nederlandse bevolking is “wit.” Dan kun je dus ook verwachten dat de overgrote meerderheid van de staf op universiteiten “wit” is. Dat gezegd hebbend: als ik om me heen kijk op de universiteit waar ik werk, zie ik medewerkers van een groot aantal verschillende nationaliteiten. Goede staf is heel moeilijk te vinden, en er zijn in het buitenland veel wetenschappers die graag aan een Nederlandse universiteit werken. Aangezien expertise bij selectie voorop staat, is het geen wonder dat het percentage stafleden van niet-Nederlandse afkomst veel groter is dan je op grond van bevolkingspercentages zou kunnen verwachten. Okay, misschien is het percentage “medewerkers met een donker kleurtje” wat lager dan de bevolkingspercentages (velen komen namelijk uit Oost-Europa, Azië, of het Midden-Oosten), maar er is geen aanwijzing dat zij minder vertegenwoordigd zijn vanwege “structureel racisme.” De studentenpopulatie immers reflecteert ook niet de maatschappelijke demografische percentages.

Tenslotte: hoezo “te westers?” Nederland is een westers land, dus mag je verwachten dat aan de universiteiten de westerse normen voor educatie gevolgd worden. Deze normen zijn gebaseerd op het wetenschappelijk benaderen van kennis. Deze benadering is dermate succesvol, dat overal ter wereld universiteiten deze aanpak volgen. Kennis wordt gedeeld over de wereld als geheel (uitgezonderd bepaalde landen waar Internet sterk aan banden is gelegd). Kortom, vrijwel alle universiteiten ter wereld volgen een “westers curriculum.” Wat is er “terecht” aan het curriculum als “te westers” beschouwen? Moeten we onze gedachten over wat feiten zijn gaan laten bepalen door religieuze inzichten, zoals we zien in bepaalde niet-westerse landen? Of wellicht door politieke stromingen, wat in bepaalde Aziatische landen gebeurt? Moeten we, zoals ik een aantal zwarte studenten heb zien verkondigen, voodoo-rituelen serieus gaan nemen?

Of slaat het “te westers” zijn op die tweede zin, dat de stand van de wetenschap het gevolg is van “wat witte mannen eeuwenlang hebben bedacht?” Wetenschap is niet wetenschap omdat het bedacht is door witte mannen. Wetenschap is wetenschap omdat het gaat over falsificeerbare feiten. Wetenschap is zelf-corrigerend. Als iemand een bewering doet zonder er argumenten voor aan te dragen, wordt de bewering niet serieus genomen. Als er falsificeerbare argumenten worden aangedragen, die ontkracht worden, wordt de bewering verworpen. Wat overblijft is een bouwwerk van feiten waarvoor de bewijsvoering zo sterk is dat we erop voort kunnen bouwen. En als er nieuwe feiten worden aangedragen, of feiten ontkracht worden, wordt het bouwwerk aangepast.

Zelfs al zou het zo zijn dat, historisch gezien, witte mannelijke wetenschappers oogkleppen op hadden en daarom de wetenschap in een bepaalde richting duwden, dan is het al heel lange tijd het geval dat wetenschap internationaal en inclusief is. Iedereen, waar ook ter wereld, ongeacht geslacht, ras, of afkomst, kan wetenschappelijke vindingen doen, bekritiseren, of onderuit halen. Een wetenschapper die een tegenargument poogt te ontkrachten door te verwijzen naar het geslacht of het ras van degene die het tegenargument brengt, wordt weggehoond en verguisd. De natuur van de wetenschap is dat het blind is voor de demografische kenmerken van wetenschappers. Daarmee getuigt de uitspraak “wetenschap is niet goed want het is gebaseerd op wat witte mannen gedaan hebben” van een dermate onwetenschappelijke houding dat degene die hem maakt slechts minachting verdient.

Kortom, het minder “westers” maken van het curriculum staat gelijk met het afbreken van de wetenschappelijke integriteit, en dat kan niet worden aangeduid als “terecht.”

In dezelfde paragraaf gaat nog een tweede alarmbel af, zij het iets minder luid dan de eerste. Aleid stelt: “Er zijn beschamend weinig vrouwelijke hoogleraren of hoogleraren en onderzoekers met een migratieachtergrond.” De alarmbel klinkt bij het woord “beschamend.” De rest van de zin is een feit. Aleid vind dit een “beschamend” feit. Mijn vraag is “wie moet zich hiervoor schamen?”

Ik heb het sterke vermoeden dat Aleid vindt dat universiteiten of de Nederlandse samenleving zich hiervoor moeten schamen. Maar waarom? Onze samenleving biedt mensen een grote keuzevrijheid, en universiteiten doen dat ook. Als je je kapot wilt werken om een hoogleraarspositie te verwerven, dan mag dat. Als je het liever rustiger aandoet, een leuke baan hebt en daarnaast gezellig met je gezin veel tijd thuis doorbrengt, of een rijk sociaal leven erop na wilt houden, dan mag dat ook. Vrijheid, blijheid. Dat die witte mannen (en een enkele witte vrouw) zo nodig statusbelust moeten zijn, hun gezin verwaarlozen, hun vrienden verliezen, maar wel veel subsidies binnenhalen en veel publiceren, is hun zaak. Dat moeten ze zelf weten, de sukkels. Zitten ze daar tot diep in de nacht op kantoor, hun gezondheid naar de knoppen te helpen, een leuk salaris binnenslepend dat hun partner vervolgens kan spenderen. Lachwekkend, maar ze willen het zelf.

Als het werkelijk een beschamende zaak is dat er relatief weinig vrouwen en weinig mensen met een migratieachtergrond zijn die carrière maken op een universiteit, dan moet die schaamte gezocht worden bij degenen die te weinig hun best doen om te concurreren met de hardwerkende carrièremakers. Maar ik persoonlijk vind er niks beschamends aan dat mensen ervoor kiezen te doen wat hen gelukkig maakt. Het is een van de grote verworvenheden van onze maatschappij dat dat mogelijk is.

Na het korte historische overzicht geeft Aleid aan dat het weinig zinvol is de samenleving in groepen te verdelen waarbij mensen met bepaalde demografische achtergronden op één hoop worden gegooid. Ze hoopt dat het niet de bedoeling is dat alle geledingen “divers” worden samengesteld, “keurig van alles wat,” want het gaat tenslotte om professionaliteit. Ze geeft aan dat de gedachte van De Graaf dat er meer smaken zijn dan “racist” en “slachtoffer” niet vreemd is (een understatement), en dat studenten per definitie gepriviligeerd zijn. Je kunt beter de grote uitval van migrant-studenten aanpakken (of de grote uitval van mannelijke studenten, zou ik daaraan toe willen voegen), of de eenzijdige samenstelling van selectiecommissies (wat ik nooit geobserveerd heb, dus ik zou daar graag eens objectieve feiten voor zien). Al met al geeft Aleid hiermee aan een redelijk gezonde kijk te hebben op maatschappelijke fenomen.

Dan gaat een derde alarmbel af bij de volgende paragraaf: “De Graaf is een tegenstander van quota voor vrouwen en minderheden, omdat ze het ‘vernederend’ vindt om ergens binnen te komen omdat je vrouw bent of een kleur hebt -– ik kan me daar iets bij voorstellen. Maar de studenten in Het Parool hebben ook gelijk als ze stellen dat ‘vanwege structurele discriminatie achtergestelde groepen vaak niet worden aangenomen, ondanks hun expertise’.” De eerste zin van deze paragraaf komt overeen met wat ik veel vrouwen heb horen zeggen, dus daar heb ik geen problemen mee. Maar de tweede zin doet mijn nekharen overeind staan.

De tweede zin is een beschuldiging van crimineel gedrag. Het is in Nederland verboden om te discrimineren op basis van demografische kenmerken. Er rusten straffen op. Dus als het objectief is dat iemand de geschiktste persoon is voor een open positie, maar niet wordt aangenomen op grond van het feit dat de persoon tot een minderheidscategorie behoort, dan is er een strafbaar feit gepleegd en moeten juridische maatregelen worden genomen.

Waar komt die bewering dat universiteiten structureel discrimineren vandaan? Zijn daar objectieve feiten voor? Zo ja, dan zou ik dit graag voor een rechtbank uitgezocht zien. Of zijn dit slechts roddels? Is er iemand niet aangenomen op een plek, en roept die persoon dat dat duidelijk vanuit racistische motieven is, zonder dat hard te kunnen maken? In dat geval kunnen we deze bewering rustig naast ons neerleggen, en stellen dat de studenten in Het Parool uit hun nek kletsen.

Maar ik vermoed dat ik wel weet wat er achter deze uitspraak zit. Het is de gedachte: “als er geen racisme zou zijn, zouden we hogere percentages mensen met een donkere huidskleur aangenomen zien worden, en omdat dat niet zo is, is dat het bewijs dat er racisme is.” Deze gedachten klinken sommigen redelijk in de oren, maar zijn het niet. Ze gaan er namelijk van uit dat er tussen groepen mensen geen andere verschillen bestaan dan alleen demografische attributen. Ze houden er bijvoorbeeld geen rekening mee dat Nederland geen homogene samenleving is waar iedereen in alles gemixt is. Demografisch onderscheidbare groepen trekken zich vaak terug in “eigen kring,” met hun eigen culturele normen en waarden. Bijvoorbeeld: moslims in Nederland worden vaak religieus opgevoed, waarbij wetenschappelijke inzichten worden afgedaan als “in tegenspraak met de Koran.” Is het verwonderlijk dat we minder moslims aan universiteiten zien dan het percentage moslims onder de Nederlandse bevolking? Dat is verklaarbaar vanuit de moslim-cultuur, en er hoeft geen verklaring gezocht te worden in “discriminatie.” Een soortgelijk verhaal kun je uiteraard ophangen over de autochtone Nederlanders die met de Bijbel in de hand geboren zijn.

Kortom, zolang er niet aannemelijk kan worden gemaakt dat racistische motieven spelen in de aanstelling van medewerkers, is het lasterlijk om te beweren dat dit toch gebeurt. Het feit dat Aleid expliciet stelt dat de uitspraak van studenten op waarheid berust, vind ik griezelig.

Het verdere verloop van Aleid’s stukje kan ik alleen maar van harte onderschrijven. Haar laatste paragraaf start met de volgende zinnen: “Natuurlijk moet de universiteit een veilige plek zijn. Ik hoop dat het voor iedereen dít is: een intellectuele vrijplaats. Een plaats waar afkomst, sekse of voorkeuren niet tellen, maar waar je die onderwerpen onbedreigd aan de orde kunt stellen.” Bravo.

Ik heb de laatste jaren in diverse landen, waaronder de Verenigde Staten, Canada, Australië, en meerdere Scandinavische landen, verontrustende ontwikkelingen gadegeslagen waarbij bepaalde universiteiten veranderen van wetenschappelijke bolwerken in instituties waar vooral gestreden moet worden voor “sociale rechtvaardigheid,” met programma’s waar geen wetenschappers maar activisten worden opgeleid. Ik heb gezien dat studenten veranderen van zelfstandige, weldenkende, kritische mensen die beseffen wat voor een gepriviligeerde positie ze innemen, in watjes die menen dat zij slachtoffers van maatschappelijk onrecht zijn, en die alle kritiek op hun gedrag afdoen als racisme, seksisme, genderisme, ableisme, of een ander neologistisch -isme. Ik zie studenten die zich gedragen als verwende kleine kinderen die door hun plaatsvervangend ouder, de universiteit, beschermd moeten worden tegen de boze buitenwereld.

Gelukkig heeft dit soort ontwikkelingen in Nederland nog niet veel aan momentum gewonnen. We zijn doorgaans een nuchter en realistisch denkend landje. Maar we gaan dit soort onverkwikkelijke zaken, waarbij eisen gesteld worden op basis van vermeend slachtofferschap, steeds meer zien. Wat de studenten van de Universiteit van Amsterdam in Het Parool zeiden, geeft me het ongemakkelijke gevoel dat ze in Nederland hopen te imiteren wat in de Verenigde Staten en Canada een aantal universiteiten aan het ondermijnen is. Het verheffen van slachtofferschap tot het hoogste goed, en het toepassen van “social engineering” om mensen macht en posities te geven die ze niet verdienen op basis van hun prestaties, moet bestreden worden teneinde de wetenschappelijke kwaliteit van universiteiten te waarborgen.

Redelijk denkende mensen met een platform, zoals Aleid Truijens, kunnen daarin een leidraad bieden door deze ontwikkelingen kritisch te beschouwen. Het feit dat Aleid termen als “terecht” en “gelijk hebben” gebruikt waar ze spreekt over sommige van de slecht-gefundeerde gedachten die de activistische studenten koesteren, kennelijk zonder het nodig te vinden enige twijfel over die gedachten uit te spreken, vind ik daarom beangstigend.


Extensions vs. careers

May 2, 2018

NWO is the Dutch organization which the government supplies with funds to distribute to scientists for their research. One of the programs that NWO runs is the “Vernieuwingsimpuls” (VI), which supplies big personal grants to scientists at different stages of their careers. The three programs in the VI are (with increasing grant sizes) VENI, VIDI, and VICI. These programs have a time limit associated with them, related to when someone got their PhD. VENI has to be applied for within 3 years after getting a PhD, VIDI within 8 years, and VICI within 15 years. These time limits are extended for biological mothers. This extension amounts to 18 months per child, for a maximum total of 5 years. The reason that NWO states for providing mothers with extensions is to allow them to spend extra time on taking care of their children.

Very recently, NWO decided to also allow the partners of the biological mothers to get extensions to the time limits, to the tune of 6 months per child. According to NWO, due to changes in societal opinions, they decided to stimulate a “more balanced distribution of professional and parental tasks between the parents” by also giving the partners an extension.

My question with this decision is: “why do partners only get 6 months, and biological mothers 18 months?” Is that really stimulating a better balance of parental tasks between the parents? If two young scientists get two kids, and one of these young scientists gets 4 years to apply for a VENI grant, and the other 6 years, who of them is under the biggest pressure to get that grant application in? Who will need to spend more time on advancing their career? The only way to ensure that these young parents can distribute their parental tasks better is to supply both of them with the same extension. NWO may think they are being progressive with their decision, but in the end they are only putting a spotlight on there being more pressure on the partner (usually the father) to advance their career than on the mother. And once one partner’s career is advancing while the other one’s career is lagging behind, when a decision needs to be taken who will spend more time at home and who will spend more time at work, it is obvious which way that decision is likely to go.

Overall, I wonder whether these extensions, especially with a length of 18 months per child, are a good idea anyway. By giving these extensions, it takes off the pressure to get grant applications in. The main way to progress one’s scientific career is to get grants. My experience is that working in science is a challenging job with a lot of tasks. There is usually little time to spend on writing a grant proposal, which may easily take months of work. There is always a high-priority task which takes precedence, and putting in a grant proposal only becomes high priority when the deadline looms. Adding 18 months per child to those deadlines basically means that a young mother is stimulated to stop advancing her career for several years.

Having a career means that you have to make certain choices. These extensions are an easy excuse to choose to put one’s career on hold, because “you get extra time to restart your career.” Unfortunately, after taking a considerable pause in advancing one’s career, it tends to be hard to get it moving again. The extensions may therefore very well be a prime cause in mothers dropping out of the rat race and deciding that they do not really need to go for the higher-level positions. Not that there is anything wrong with deciding that a nice balance between home life and work is preferable over sacrificing one’s home life for a career. That is a perfectly valid choice. My problem is that the extensions which NWO provides, which are evidently meant to allow mothers to have a home life without sacrificing their careers, may actually stimulate mothers to give up on their careers.

Moreover, I wish to point out that there are many reasons why someone’s work life may suffer from a set-back. Wanting to stay home with a new-born child for a while is only one of them. But how about going through a divorce? Changing jobs? Having to take care of an ailing family member? Being a single parent who is not the biological mother? Wanting to take a sabbatical? Needing health recovery after an accident? Most people with a job find that they sometimes get into personal situations which need them to put career advancement on hold for a while. What makes young mothers such an exception that they should be richly compensated for choosing to stay home with their children?

I am definitely not against time limits on submitting grant proposals. Time limits ensure that people submit these proposals at the “right time” in their careers, and, in particular, heed people against giving too much priority to tasks other than writing strong grant proposals, which harms them in the long run. However, automatically extending the time limits for specific groups of people is rather arbitrary and may very well be harmful to the careers of many of these people.

I am not opposed to giving people extensions to the time limits, but I would not make them automatic (and not this long). Simply let people apply for extensions and let them state their reasons for wanting one. This takes off the arbitrary edge, while also putting up a barrier for making use of extensions. Such a barrier is a necessity, as it is a harmful illusion to think that these extensions allow a parent to have a full home and social life without damaging their career.


Killer robots are here already

August 25, 2017

At the International Joint Conference on Artificial Intelligence 2017, an open letter was released, signed by over one hundred top scientists and industrialists in artificial intelligence, calling for a ban on the development of autonomous, artificially intelligent weapons, often referred to as “killer robots.”

This vapid gesture is equivalent to calling for a “ban on the development of knives that can be used to murder people.” The problem is that almost any device that can be taught behavior and is allowed autonomous functioning, can be employed as a “killer robot.” And all industrial artificial intelligence research advances intelligence, learning ability, and autonomy of machines.

Elon Musk might be in favor of a ban on the development of killer robots, but his Tesla company works on autonomous self-driving cars. Recent terrorist activities have demonstrated how cars can be used as weapons. You only need to teach a car to hit people instead of avoiding them.

Mustafa Suleyman might want to stop research into killer robots, but at the same time his DeepMind company is the leader in deep-learning research, which aims at allowing machines to learn patterns and respond to them. Such pattern recognizers can be easily placed in smart missiles or weaponized robots to autonomously find viable targets.

Jerome Monceaux signed the letter, while simultaneously heading Aldebaran Robotics, which develops general-purpose robots which can be taught or programmed to do anything — including using weapons and going on a murder-spree.

And the list goes on.

The whole point of artificial-intelligence research is to allow machines to do things that humans can do, preferably more efficiently and effectively, and preferably with a high degree of autonomy. Moreover, almost all modern artificial intelligence research is based on machine learning, i.e., teaching machines to behave in a particular way rather than directly programming them. Consequently, almost any artificial intelligence research can be used to teach machines to help people, or to behave as a weapon. This entails that machines that have the ability to operate as killer robots already exist.

Basically, the call for a ban on the development of killer robots amounts to a plea along the lines of: “Look, we are developing all this great technology which will bring fantastic benefits to humanity but please, please, please do not use it to murder people.” It is a call for sanity on the part of governments, the military, and terrorist organizations so that they won’t use the technology for evil. And we all know that the sanity of governments, military, and terrorists varies.

You cannot stop the possibility of (further) developing killer robots without a world-wide halt on artificial intelligence research altogether. I do not think that that is what any of these people who signed the letter, or anyone else, really wants. Or that it can be enforced, for that matter.

The best you can do is realize what artificial intelligence can be used for and then build in protections against misuse. For instance, autonomous self-driving cars should be strongly guarded against attempts to reprogram them. This is in the hands of Elon Musk and his competitors. Rather than calling for some kind of ban, they should do their jobs properly. And while I think they are trying to do a proper job, their call for a ban sounds like them trying to place the responsibility for misuse of their technology in the hands of others.

Any technology can be misused, and usually is. That is no reason not to develop beneficial technology. The benefits of autonomous artificial intelligence can be great. The dangers of it are lurking in the autonomy — technology which allows machines to operate autonomously, taking autonomous decisions on how to act, should be surrounded by stringent safeguards against the machines taking harmful decisions. But probably the biggest danger is not in the artificially intelligent machines themselves, but in the humans who place unwarranted trust in them to take autonomous decisions.

I applaud the fact that many influential people consider the dangers of artificial intelligence research seriously. The call for a ban, however, sounds like an after-the-fact plea.


Diversity VI: Normal discrimination

August 22, 2017

As a final follow-up to my previous post on the issue of enforced diversity in the workplace, I wish to expand on the criticism that I gave on those scientists who deny the notion that biology has an influence on the different roles of men and women in society. The main argument that is given by these scientists is that the biological differences between the sexes, while existing, are smaller than is generally assumed. From this notion they derive that, based on biology, you should see only small differences between behaviors of men and women. Since this is not what can be observed, they conclude that men and women are treated differently because of (only) cultural influences.

These scientists do not understand the normal distribution.

We may assume that most characteristics of men and women, both physical and mental, are approximately normally distributed. For the sake of argument, let’s say that due to biological causes, men on average have a slightly higher aptitude for abstract thought than women, while women on average have a slightly better social sense than men. If you then give a large bunch of men and women an IQ test (which measures abstract thinking) and an EQ test (which measures social thinking), then perhaps on average men have an IQ of 100.5 and women an IQ of 99.5, while men on average have an EQ of 99.5 and women an EQ of 100.5 (I am pulling these numbers out of thin air, they are just examples). The 1-point difference shouldn’t be noticeable in everyday behavior.

However, the 1-point difference is for the average. At the extremes of the normal distribution, the differences are much more noticeable. If the shapes of the Bell curves for men and women are the same, but with a difference of 1 point for the averages, and we look at the 5% highest scorers for both groups, the average of them differs a lot more. Looking at it in another way, if you are only checking those with an IQ or EQ higher than 150, the 1-point difference for the average can easily translate to 80% of the people with an IQ higher than 150 being men, and 80% of the people with an EQ higher than 150 being women.

The fact that companies and institutions want to hire “the best” for crucial positions, entails that they discriminate to look for the high-end tail of the Bell curve which describes the population distribution for the traits which they need the most. Even if on average there is little difference between the aptitudes of men and women to do particular jobs, when you discriminate to get the best, you will generally end up with a severely skewed distribution of the sexes.

This effect is magnified by the fact that for many traits, the variance differs for the sexes. In particular, it has been reported that men tend to have a higher variance than women. This means that men fit a flatter Bell curve than women for many traits, which translates to there being a denser population of men at both extremes, even if there is no difference between the averages. In layman’s terms: if in a particular area there are more genius men than genius women, then this is offset by there being more moronic men than moronic women (though there will, of course, still be both genius women as well as moronic women).

Even if the aptitudes of men and women for a particular job are the same, fitting the exact same Bell curve, then there is still the fact that the interest for them to do that job may differ. Again, if interest in tech is normally distributed over the population, and there is on average little difference between the interests of men and women for tech (which might be the case: most people are not really interested in tech jobs), then you might still find big differences between the number of men and the number of women who are highly interested in tech, i.e., whose interest is at the high end of the Bell curve. Which means that even if women are equally able to do a tech job, that does not mean that you will find many women actually preferring to do it over some other job.

Note that the same may be noticed in high-end jobs which see a dominance of women, such as medical specialists. At a time when few women entered the job market, physicians were usually male. Nowadays, when about 40% of jobs are held by women, physicians are predominantly female — either because being a physician requires a skill set for which the best candidates tend to be women, or because it is a job that women have, on average, a higher preference for than men. Regardless, being a physician is a high-end profession where women rule — which makes mincemeat of the notion that some kind of male oppression is keeping women out of high-end jobs.

Enforcing an equal distribution of the sexes for those positions for which high skills and/or high motivation are needed, leads to either lowering the quality of how the work is done, or appointing people who would rather be doing something else.

By the way, the discussion above does not mean that there actually are only small biological differences between the sexes. In my previous post I gave three reasons to suspect that the biological differences are quite significant. However, even if you assume that there are only small biological differences between men and women, distribution of the sexes in high-end jobs will usually be unbalanced.


Diversity V: Nurture vs. nature

August 11, 2017

I feel the need to make a few more statements about the firing of James Damore by Google, which I discussed before, in particular insofar it concerns his claim that biology may be partially responsible for observed differences between the interests of the average man and average woman in society, and how this claim seems to be interpreted.

As Google CEO Sundar Pichai wrote on the reason for Damore being fired: “to suggest a group of our colleagues have traits that make them less biologically suited to that work is offensive and not OK.” I read Damore’s treatise, and nowhere in it I can find him saying that his female colleagues are less biologically suited for their work — he only states that the biological differences between men and women may make the average woman less suitable for tech and leadership than the average man. He is not saying that his female colleagues are less suitable than his male colleagues. He only says that these biological phenomena may explain why we see less women in tech and leadership, and that trying to enforce an equal representation of the sexes in tech jobs in Google may be misguided.

The only way that someone can interpret Damore saying something about his colleagues is if you assume that what he says about the average woman holds for all women — and he quite clearly says the opposite; he even spends a considerable portion of his treatise discussing that it is a mistake to regard each individual as representative for their sex. That, however, is what is done when you treat women differently than men just because they are women (which is what Google does). You might think that a CEO who fires a person would at least check whether the reason that he gives for the firing holds water, but it looks like Damore’s firing was not because of what he factually said, but because many people found him offensive — the numerous tweets and responses on the Internet suggest as much.

So what could people find offensive about his text? It cannot be the statement that people should not be equated to the average of their sex — only the opposite could be found offensive. Considering Pichai’s statement, and what I read in many of the comments on Damore’s text, what is found offensive is Damore’s suggestion that the cause for observed behavioral differences between men and women may be partially found in biology.

If you examine literature, you will find that there are certain scientists who state that, while there are definitely biological differences between men and women, they are much smaller than generally assumed, and that all behavioral differences between men and women are the result of culture. If you believe that what these scientists say is undeniably true, then you may feel offended by anyone who is merely suggesting that there are innate differences between men and women. In comparison, if you believe that the Bible speaks holy truth, then you will probably feel offended by anyone who says that there are indications that the Earth is 4.5 billion years old. Such a person is an infidel and should be burned at the stake.

However, what these scientists say is not scientific consensus. Other scientists state that, while culture obviously has an influence on behavior, there are clear, experimentally demonstrable, indications that the innate biological differences between men and women influence interests and behaviors, which partially explain the differences that we observe between the averages of men and women on a population level. Even without running any experiments, I’d say that there are numerous easily-observable indications that biology influences behavior. I give three of them:

  • In most animals, mammals in particular, there are clear behavioral differences between the sexes. Animals do not have cultures, so these differences are all biologically driven. I see no reason why humans would be exempt from these biological influences.
  • Biological differences between men and women result in the production of different quantities of particular hormones. We know that hormones not only influence physical development, but also behavior. In particular, in puberty, when the hormonal production starts to get radically different between men and women, the behaviors of men and women start to diverge extensively. Moreover, if you feed people particular sex-related hormones, their behaviors change too. Again, this shows that biology influences behavior.
  • In western countries, people have been given a lot of freedom of choice. In north-western Europe, the cultural pressures on men and women to stick to particular roles in society have been mostly eradicated. If the theory that all behavioral differences between men and women are the result of culture would be correct, the differences in interests between men and women would dwindle away. This has not happened. Women have entered the job market, but on average choose more people-oriented jobs than men; for instance, 50 years ago there were many more male physicians than female physicians (as few women had a job), but these roles are reversed now. However, 50 years ago there were many more male engineers than female engineers, and these roles have not changed. Also, 50 years ago it was not common to have a part-time job for men or women, but nowadays women predominantly work in part-time jobs, while men do not. If release of cultural pressure increases differences in behavior, that is a clear indication that the differences are at least partially caused by something else than culture.

Moreover, the arguments that I read on why some people believe that only culture plays a role in determining differences in interests between men and women seem faulty to me. For instance, it is pointed out that the roles of men and women are not innate, as there is an African tribe that has a matriarchal structure in which women have the role that is traditionally associated with men, and vice versa. To me, it seems that the existence of such a tribe is the exception that confirms the rule: everywhere in the world, one tiny place excepted, men take a particular role and women another, which shows that these roles are nearly universal. The fact that the exception exists only shows that culture may overrule the traditional roles. Now, I am not an expert on this topic, and thus I might have missed some very convincing arguments for assigning all responsibility for behavioral differences to culture. However, I think that if convincing arguments would exist, the scientific world would have reached consensus by now, and that clearly is not the case.

The claim that only culture influences behavior is much, much stronger than the claim that behavior is partially the result of culture and partially the result of biology. If you want to make the first claim, you have to come up with strong evidence that biology does not play a role. I have not seen that evidence. And since there are no ethical experiments that you can perform to exclude biology as a partial cause for observed differences between men and women on average, it will be very hard to come up with it.

I am not saying that it is definitely false that only culture is responsible for behavior. It might be true, though considering the observations that we can make, it is unlikely to be true. The problem is that just like you cannot prove that biology has no influence on behavior, you cannot disprove that only culture is responsible for behavior. However, claiming that biology has no influence means upholding a belief in the face of observable facts that point in a different direction. It is a belief that has to resort to shaky claims like “unconscious biases” to explain why certain observable facts exist. And since it is hard to use factual arguments to defend this belief against grounded criticism, believers may feel a need to attack the credibility of non-believers by accusing them of being reprehensible sub-humans, i.e., “sexists.”

James Damore’s statements seem innocent to me, but to someone who upholds an ideology that is based on the premise that there are no biological differences between the sexes, they are dangerous, and must be attacked. Labeling him a “sexist” is an easy way of making him harmless.


Diversity IV: Google ideology

August 8, 2017

Senior Google engineer James Damore wrote a memo criticizing the Google mantra that the underrepresentation of women in tech and leadership is solely caused by biases. He poses that there are scientific indications that biological traits may make women on average less interested or even less suitable for such roles. Note that he is very careful in underlining the “on average,” explicitly stating that there is an enormous overlap between the populations of men and women in all their attributes, and that people should not be assessed as the average of their groups, but on an individual level. He also explicitly states that he does not deny that sexism or biases exist, and that he values diversity and inclusion. His main issue is that he found that he is working in a psychologically unsafe environment, since even suggesting that anything but biases are the cause for different representations of men and women in the tech industry is cause for shaming and misrepresentation, and risks being fired.

Google promptly confirmed his statements by firing him for “perpetuating gender stereotypes.”

Contrary to many of those who accused James Damore of sexism, I read his 10-page document and examined the links he included (which, unfortunately or maliciously, were removed by Gizmodo when they republished his memo). It is a well-spoken treatise, in which the only sexist statement that I could find amounts to the suggestion that there are biological differences between men and women, and providing scientific support for that statement. I know that in many circles such a statement is considered sexist, but that means that in these circles scientific facts are considered sexist.

I do not agree with everything that Damore states, as I think some of the underlying science is oversimplified. But the point is not whether Damore is correct in everything that he says. The point is that bringing up viewpoints which challenge the reigning ideology should lead to an open discussion, not be a reason for getting fired.

There is overwhelming scientific evidence that there are biological differences between men and women (it feels weird to even have to make such a statement — I mean, most people know how sexual reproduction works, right?). There is evidence that these biological differences lead to some innate differences in interests and aptitudes between men and women on average (emphasis on the “on average”). Even if you wish to marginalize the biological causes for the average differences in interests and aptitudes between men and women, these differences can be observed. That means that yes, the stereotypical man differs from the stereotypical women. It does not mean that every person is like the stereotype of their sex. It just partly explains the differences that we can observe on a population level. To what extent are these differences caused by biology, and to what extent are they the result of cultural biases? We don’t know. And if we are not allowed to have a discussion about it, we will never know, and we can never come to diversity programs which tackle the biases in an effective manner.

I have argued before that some of the enforced diversity programs that I have observed are misguided. I have been cautious in expressing that opinion, as I know it can lead to fast accusations of being sexist, which is a label that you better avoid. However, I still felt sufficiently safe in expressing that opinion as I think that I am working in an environment where people are at least willing to examine and discuss evidence rather than condemn automatically. Having to work in an environment where expressing a well-founded idea can lead to getting fired seems horrifying to me.

Science stops where ideology takes over. It seems to me that a company such as Google, which is founded on technical and scientific research, is doing itself a great disservice by trying to silence those who hold a contrary opinion or are trying to challenge ideas.


AI storytelling

August 6, 2017

Recently I read in a newspaper a list of average predictions of AI researchers on when certain achievements in AI would be reached. There were several predictions for the coming 5 to 10 years, such as an AI winning a game of StarCraft against a human champion (2022), and composing a top-40 song (2027). Only one prediction was made a considerable number of years in the future, namely writing a New York Times bestseller (2050).

I was not surprised about the short-term predictions. These were all straightforward extrapolations of today’s research. For instance, a lot of time is invested in creating StarCraft AI, and we know that a computer already has a huge advantage over humans in its speed; it just needs to get a bit better tactically to defeat human champions. Similarly, computers already write music that is indistinguishable from what humans compose, so I can see a computer write a top-40 hit today — the main problem I see for writing a song that lands in the top-40 is that the quality of the song is only a very small factor in determining whether it becomes a hit.

Why is writing a bestseller considered to be much more difficult than any of the other AI tasks?

Writing a novel is very different from composing music or creating a painting. When listening to music or watching a painting, people give their own interpretation to what they hear or see, and the computer can get pretty far by simply recombining elements of music or paintings that it has been trained with. For instance, David Cope’s first attempts at letting a computer compose music amounted to hacking Bach’s sonates into measures which were stored in a database, and then recombining these measures by making sure that the last note of each measure was the same as the last note that originally came before the measure that was chosen next. This resulted in thousands of sonates which sounded more or less like Bach sonates. The computer did not need to understand what it was doing. In contrast, when writing the text of a novel an AI needs to understand what it is writing, otherwise the text will not make sense.

You might think that no real understanding is needed to create a text, as we already see human-readable text produced by computers today. In particular, newspaper articles are often written by computers. Examples are weather reports, stock market analyses, and sports reports. However, in these cases the computer is not really producing an original text. The computer simply gets the data that need to be reported (temperatures, market fluctuations, goals scored) and translates the data according to specific rules into a text. No creativity is needed. To produce a novel, the computer must come up with an original, sensible plot that has relevance to humans, and turn that plot into a captivating text. As far as I can see, the computer cannot do that without quite deep understanding of the human condition, human emotions, human language, and the human world. And at the moment, we have no idea how we can give a computer such understanding.

Someone who has heard of “deep learning” might think that it is sufficient to train a computer with existing novels to allow it to produce a new novel. But what are you then really training the computer with? You are training it with strings of words. This might lead to a computer being able to recognize that certain strings of words are likely to be sensible sentences, but not that a string of 40,000 words is a sensible, well-readable novel. Examining words does not equate examining the plot, the meaning, or the literary quality of the novel.

You might think that we may solve the problem of making the computer create a bestseller by simply letting it produce random texts, and then assess the quality of the texts with respect to them being a bestselling novel. Using an evolutionary approach, this might reasonably quickly lead to a novel that scores high on bestselling quality. This approach might actually have merit to it, if we could give the computer an algorithm that rates a text as a bestseller. We do not have that, as even humans cannot predict whether a novel will be a bestseller.

Take, for instance, the first Harry Potter novel, which was rejected by virtually all British publishers, and only produced in a very small quantity by the last one because his 8-year-old daughter liked the book. Considering Rowling’s unskilled writing and weak plot construction, it is not surprising that the publishers did not see her combination of a mid-20th-century boarding school novel with a childish version of Lord of the Rings as likely to succeed. Expert humans did not assess Harry Potter to be the commercial success that it came to be. So expert humans cannot teach a computer to do it for them.

If expert humans cannot tell a computer how to rate a novel, you might still envision an approach by which a computer determines by itself an evaluation function for bestselling quality. If you have millions of books which are all labeled with their relative sales figures, and extra data with respect to the time when and place where the books were a success, you may be able to use them to train a computer to come up with an evaluation function that can accurately predict from the contents of a novel whether or not it will be a success. Perhaps that is possible. Perhaps not. Frankly, I think that if it was easy, then all those publishers who rejected Harry Potter would have internalized an algorithm like that and would at least have seen some value in the book, but evidently they did not.

If a computer would have a much deeper understanding of the world than any human has, it would have insights that humans cannot have. And with such insights, be able to predict bestselling quality. I believe that in principle it is possible for a computer to have much deeper understanding of the world than humans have, but we are far, far away from having such a computer since, as far as I know, nobody has any idea on how to give a computer what is needed for it to gain understanding. The conclusion that I must draw is that it probably is not impossible to get a computer to write a bestseller, but that creating a computer that can do that is not a straightforward extrapolation of the state-of-the-art in AI. Therefore, attaching any year to it is unwarranted.

So where is the year 2050 coming from in the minds of the average AI researcher? I think it represents “About 25 years in the future? Who knows what we can do by then!”

Basically, predicting an AI achievement for 2050 is equivalent to AI researchers saying “we have no idea.”


Ethical cars

June 28, 2017

The first completely autonomous machines that will invade society as a whole might very well be self-driving cars. With “completely autonomous” I mean that these cars will perform their duties without any interaction with their owners, making their own decisions. Obviously, there is great commercial value in such transportation devices. However, allowing them to take responsibility for their own actions in the real world may involve considerable risk. For how can we be ensured that the decisions of these cars are in alignment with what we humans find morally acceptable?

A typical scenario that I get confronted with, is a self-driving car which has to swerve in order to avoid hitting a dog, but if it does that, as a consequence, hits a human. While obviously we would prefer the car to avoid hitting both dogs and humans, if there is no choice but to hit one of them, we would like the car to then choose to hit the dog. A potential solution to this scenario would be to outfit the car with ethical rules along the lines of Isaac Asimov’s three laws of robotics, e.g., with a rule that says “do not harm humans” given priority over a rule that says “do not harm dogs.”

However, the specification of such rules is not a trivial matter. For instance, it is logical that a rule would state “you have to obey the laws of traffic.” This would entail that the car is not allowed to drive through a red light. But what if the car stops for a red light, while a traffic warden motions it to continue driving? You may update the rule stating that an exception is made for directions given by traffic wardens. But what if there is no traffic warden, the car has stopped for a red light, and a police car sounding its horn is coming from behind and cannot get past unless the car drives forward a bit (through the red light) to park to the side? You may update the rule even more to take that situation into account, but is this then covering each and every situation in which the car is allowed to break the rule that it should stop for a red light? Probably not.

The point is that human drivers ever so often break the rules of traffic to avoid a problematic situation. You are trying to pass another car which drives fairly slowly, and suddenly that car speeds up. You can still get past, but you have to drive faster than the speed limit for a few moments. So that’s what you do. Or you are at a crossing in a deadlock with two or three other cars. One of them has to break the rules and start moving, otherwise they will all be stuck there forever.

The point is that human drivers improvise all the time. They know the traffic rules, they have been trained to recognize safe and dangerous situations, and know how to anticipate on the behavior of other drivers. And sometimes they bend or break the rules to avoid problems. A self-driving car that cannot improvise is dangerous. However, a consequence of the need for improvisation is that any rules that we would want to impose on the car, it should be able to break. The only alternative would be to envision each and every situation in which the car could find itself and specify the exact behavioral rules for it to deal with all those situations. Clearly, that is impossible.

So how do we get a car to behave like a responsible driver without laying down an endless list of rules? The answer is: by training it. First, we let the car drive in a highly realistic simulation, punishing it every time that it causes a situation that is undesirable, and rewarding it when it manages to perform well. A learning structure can incorporate the lessons that the car learns, thereby bringing it ever closer to being a model driver. Once it is perfect or almost perfect in the driving simulation, it can be let loose on the road under the guidance of a human, continuing learning. In the end, it will behave on the road as well as, and probably a lot better than, a good human driver.

How will such a car deal with a choice between hitting a human or a dog? It is likely that similar situations will have cropped up during the training process — maybe not with exactly the same race of dog and the same human as in the real situation, but as the car has been trained instead of having been given specific rules, it has the ability to generalize, and it will make the choice that is closest to what the training would have rewarded while avoiding choices that the training most likely would have punished. In other words, it will choose to hit the dog to avoid hitting the human, just as it would likely hit a cat, a moose, a badger, or a duck in order to avoid hitting a human.

It might, however, in a situation where someone pushes a mannequin in the road, hit a dog to avoid hitting the mannequin — not because it thinks the mannequin is a human, but because the situation of hitting the mannequin more closely resembles hitting a human than the situation of hitting a dog does. If we do not want the car to make that choice, we should ensure that its training regime includes situations in which it has to deal with objects that resemble humans but are not humans. This, however, could lead to a situation in which it chooses to hit a highly inert human to avoid hitting a dog. That’s the problem with allowing a car to make its own choices based on how it is trained: you can probably always find an exceptional situation in which it is not doing what we hoped it would do. The same is true for humans, of course, and in the end the self-driving car will probably still be a much safer driver than any human.

So if one wonders how we can be sure that the ethics of a self-driving car will be acceptable to us humans, the answer is that we can only draw conclusions based on observations of how the car deals with tough situations. We will not be able to open up the car’s brain and examine some kind of ethics module to read how it will deal with situations that come up. Therefore there is no way for us to be “sure.”

We can only draw comfort from the fact that if at some point the car takes a decision that we find doubtful, we can punish it and it is likely to make a different decision when a similar situation comes up again. It will be less stubborn than the average human in that respect.