Scientist specialised in security for machine learning blow the whistle : "Google poisoned the science community, which now amplifies its disinformation"

Dr. Lê Nguyên Hoang is a french computer science researcher in IA who taught ethics of AI on the internet (among other thing). In his last video, the researcher specialised in security for machine learning blow the whistle on practice in the academical world affecting privacy and security for billions of person around the world.

In an hour long video, he speaks about how articles containing volontary important ommissions and disinformations have take increasing space in the academical references of AI research, how articles he cowrites to contradict such articles, with proofs, were belittled and refused by important scientific journals that published the false articles in the first place.

In every example find of misinformation one of the authors was a Google employee.

The ton became more and more explicitly accusatory as we progress though the video :
He starts by telling that the affirmation, non-based on anything, that federated learning is more privacy-respecting than traditionnal centralised learning, is not ; simply because its point is to find the same result than a centralised learning method.
Then he explains that the mathematical proof that the performance and privacy in machine learning are incompatiblable was found by his colleagues and him but it is yet to be published as it was refused for not telling the same thing that the main-stream articles that made the hypothesis it is true.
He continues, saying a that generalisation of data is a bad thing in itself when applied to every field. For the algorithms deployed by tech geants, the main source of data is people expressing themself freely, on social media for example. Such, source data contains racisms, intolerace and other misbehavior that, for obvious moral, reason should not be generalized or reproduced.
At some point, he even accused the author of a paper, employee at google, to delibaratly propagate dangerous AI method that are today cause of escalation of violence, hatred leading to already happening genocide.
He accused Google of volontarily making up scientific proof that their work is safe and should keep developping while turning research into a field motivated by monnaie, forcing scolar to publish frequently about trending topics.

The video is truly interesting for people knowing AI and non tech-savy people. It gives a lot of resources, for privacy, security and ethics in IA in the forms of YouTube videos and scientific paper.

“Google poisoned the science community, which now amplifies its disinformation” by Dr. Lê Nguyên Hoang on Science4all (english) :

“This video debunks 4 dangerously misleading scientific papers, all with a Google co-author, that denigrate the urgent need for massive audits and investments in security and ethics. Terrifyingly, the scientific community has scientifically validated and massively propagated what appears to be intentional disinformation.”

Do you think this kind of warning can become viral and wake up people ? Or will it be like the rest and don’t be known ouside the community of people already convince by the importance of privacy ?

2 Likes