Kate Crawford reminds us that the revelation of these ethical scandals first changes the technical processes before the working conditions – like what happened at the beginning of the 20th century during the first revelations about meat slaughterhouses : this pushed to legislate on food security, not on the condition of exploited workers. ” The persistence of this pattern highlights how power responds to criticism: whether the product is cow carcasses or facial recognition software, the response is to accept margin regulation, but leave the logic of underlying production, writes Kate Crawford.
An inhumanity at work which also draws its foundation from the fact that tech is carried by the workaholic young engineers – mostly men without families who praise the fact of working 24/7…. With the notable difference that they can expect to make fortunes with it. In the realm of inequality, tech is king.
The next chapter discusses the supposed intelligence of data. And Kate Crawford to point out the gigantic problem of an extension of this data trade without ethical reflection: “ the NIST (National Institute of Standards and technology) databases prefigure the emergence of a logic that has now invaded the entire tech sector: the absolute conviction that everything is given usable. It doesn’t matter where the photograph was taken, whether it reflects a moment of vulnerability or suffering, whether it humiliates the subject. It has become so normal in the industry to take and use whatever is available that few people question the underlying policies. “.
And the chapter to tell more edifying anecdotes than the others, like the fact that the first training data extracted from private emails of executives of Enron (the consulting company at the origin ofa resounding bankruptcy) were seized by the courts, on the grounds that ” the public’s right to the truth trumps the individual’s right to privacy – these exchanges found their way into thousands of scholarly articles. RIP privacy.
The data always claims to seek neutral information, but the object of the query always responds to a precise motivation. Thus, with a database listing the movements of New York taxis, we can discover which celebrities frequent striptease clubs, but also which drivers are Muslims by checking the coincidence of their breaks with prayer times. Above all, these raw data are worth nothing and if “data is the new oil”, only the refined interest companies, ie classified data.
The chapter devoted to classification by AI is undoubtedly the most staggering, as we see how political, human, biased and ineffective these classifications are. To tell the truth, we are bewildered by the amateurism and the crass stupidity of those who call themselves demiurges of artificial intelligence. Kate Crawford starts from the gigantic ImageNet project, which intended to classify all the people for whom we have photos in an extension of filing as has been done since the 19th century with identity documents. But the way to store the found photos is bewildering: beyond “woman” and “man” and “Caucasian” or “Asian” – which is already problematic in itself – the promoters of the database have listed the images in classy “dumb”, “whore”, “failed”, “inclined”… Under pressure, the team had to delete 1593 of the 2832 categories that they considered “dangerous” and deleted 600,000 images… A straw.
The absence of ethics, of human reflection in the classification poses enormous problems in the biases of algorithms, well-documented evils: discrimination in employment reinforced with offers sent primarily to white men for finance jobs, to women for care… All the human blemishes multiplied by machines and wrapped in a misleading image of “technological neutrality”.
A single mantra: “illimitism”
The last chapter concerns the use of AI by States with public-private partnerships mixing the military field and mass surveillance for more or less avowable and more or less dramatic ends depending on the regimes – from the scandal of the NSA eavesdropping by Americans at the Orwellian social credit system Chinese. We meet Palantir Technologies and DARPAgigantic private projects, often libertarian, on which the public powers rely, all shame drunk.
The fascination of the machine on the public could lead to the absence of human decision-making in areas as sovereign as justice. And it seems to have no end, since the mantra of AI is “illimitism” – to use this neologism coined by the researcher John Chapoutot when he talks about the doctrine that feeds Jeff Bezos and Elon Musk.
Moreover, the mischievous coda of the book shows that after having defiled the planet like no other before them, the two paragons of AI now wish… To conquer space.
We wish to say thanks to the writer of this write-up for this awesome web content
“The only conviction of tech champions is that all data is potentially exploitable”
Check out our social media accounts along with other pages related to themhttps://www.ai-magazine.com/related-pages/