Amnesty International has warned that artificial intelligence, often promoted as a security or anti-discrimination solution, actually exacerbates racism and other forms of discrimination and undermines fundamental human rights. There is.
“Artificial intelligence is touted as a necessary safety measure and an effective solution to combat discrimination. But far from solving these problems, artificial intelligence is amplifying racism and discrimination, It perpetuates rights violations,” said Katya Lu, Amnesty International's technology and human rights officer. International, April 24, during the release of the annual “State of Human Rights'' report.
These risks are heightened by two main factors. One is that it is a critical election year in which more than half of the world's population will be asked to vote, and the other is that the international balance has been disrupted by a number of violent conflicts in places such as Gaza, Ukraine and Sudan. This volatile landscape benefits AI, which thrives in the absence of strict regulation and fuels these tensions.
Discrimination facilitated by algorithms
According to Amnesty International, different AIs can have different impacts depending on their use, but may contribute to racial or ethnic discrimination. “For example, during the 2020-2022 conflict in Ethiopia, the conflict in Myanmar, and the Rohingya conflict, we observed that Meta (Facebook) algorithms amplified violent content that incited hatred against Tigrayans. “The focus on engagement with violent content increases hatred,” Lu said.
Beyond digital giants’ algorithms, new technologies such as generative AI, which can create content independently of simple commands, are also disrupting political life. They don't shy away from democracy either. For example, in the United Kingdom, where elections are scheduled for January 2025, a series of “deepfakes” (artificial intelligence-generated images) were distributed showing Labor MPs and opposition leader Keir Starmer insulting employees. , disrupted politics in new ways. .
A similar incident occurred in the United States during the New Hampshire Democratic primary in January. Months before the November presidential election, Joe Biden's rival's campaign manager admitted that the US president used AI to create a call urging voters not to vote against Biden. Ta. The fabricated audio disrupted the early primaries.
The latest Amnesty International report says facial recognition and emotion detection tools, also forms of artificial intelligence, are further contributing to unequal treatment, even when used by governments.
Europe, for example, was criticized by Amnesty International for using “iBorder Control” software that acts like a lie detector during border checks. The system, used at the borders of Latvia, Hungary and Greece and funded by the European Union, is said to encourage discrimination, NGOs claim.
“Borders are interesting for governments because they can test tools on vulnerable populations,” Lu said. “iBorder Control, like others, is labeled as a 'neutral' technology that simply assists in human decision-making, but it is designed by humans who can choose their criteria and targets specific groups. This is not the case, as it can be exploited to
Missed opportunity with AI law
According to Amnesty International, the widespread use of artificial intelligence for security issues remains largely unregulated, although its effectiveness and necessity have not been proven. But on March 24, the European Parliament adopted the Artificial Intelligence Act, a regulation that establishes rules for artificial intelligence across all areas, including algorithms, security and creative. Lu said the regulation “doesn't go far enough. It's the first in the world to have a regulation that addresses all of these points, and some states have started ambitious projects, especially when it comes to facial identification.” It's a shame they didn't promote it.” ”
As the Olympics approach, France becomes the first country in the European Union to legalize the use of artificial intelligence for surveillance purposes. The move concerns Amnesty International, despite assurances from the Home Office that its use should be limited in time. Amnesty International is concerned that such systems could become permanently entrenched in everyday life.
