<trp-post-container data-trp-post-id='25538'>Intelligence artificial : can we the tax from racism ?

Recently, Youth Laboratoriesa Russian start-up led by 28-year-old entrepreneur Alexey Shevtsov, has launched Beauty.AIThe first beauty contest to be judged by artificial intelligence: a particularly original idea, which only ended up unleashing the vindictiveness of Twitterers and other socialites against it on the grounds of... racism! And with good reason: none of its winners had dark skin!
Youth Laboratories wanted to use the buzz created by this competition to publicise its mobile application designed to use AI to monitor the progress of wrinkles and, above all, the effectiveness of its treatments to slow down skin ageing. The opposite effect was guaranteed, as the start-up was immediately accused of manipulating the codes to favour white candidates!
These are serious allegations, the foundations of which appear complicated to verify, as no-one has access to the multiple algorithms used - but it doesn't matter! However, two or three conclusions can easily be drawn, without necessarily seeking to accuse the directors of Youth Laboratories of racism.
The first is that AI is not yet a perfectly mastered discipline: remember that Tay, the artificial intelligence to which Microsoft had offered a Twitter account, had already suddenly become violently racist, proclaiming "Hitler was right, I hate the Jews" - to explain this slip, a few twitterers had gently drawn it onto a slippery slope...



The second is that AI is nothing more than the result of human programming - at least, not yet; tomorrow it will be AI that creates new AI.



It's clear that the programmers behind Beauty.AI were more partial to white girls than others. Not unlike those at Northpointe, who developed software to help American courts in several states decide whether to release defendants: with them, it was better not to be coloured!



The last is that relying on artificial intelligence alone to make strategic decisions presents serious risks. When humans argue with each other, every argument can be called into question at any moment; when a manager uses a highly sophisticated tool based on AI - and which will often have required substantial investment - he will all too often tend to adopt the very comfortable attitude of sheltering behind the 'scientific' aspect of the process... without even remembering that his tool also contains all the subjectivity of those who created it.

They we have fact confidence. Discover our achievements