Amnesty International’s recent report presents a strong critique of artificial intelligence (AI) and highlights concerns regarding its growing involvement in various forms of discrimination that contribute to ethnic, racial, and social tensions.
The organization warns that despite being touted as a solution for security and anti-discrimination efforts, AI exacerbates racism and other discriminatory practices, thereby undermining fundamental human rights.
Katia Roux, advocacy officer for technology and human rights at Amnesty International, emphasized during the presentation of the annual “State of the World’s Human Rights” report on April 24 that AI technologies, often hailed as effective solutions, actually amplify racism and discrimination, perpetuating violations of fundamental rights.
These concerns are magnified by two primary factors: a significant election year, with more than half of the global population expected to vote, and ongoing high-intensity conflicts in regions like Gaza, Ukraine, and Sudan. In this volatile environment, AI thrives due to the absence of strict regulation and contributes to escalating tensions.
Algorithms Driven Discrimination
Amnesty International highlights that various AI applications can lead to racial or ethnic discrimination, with different impacts depending on their use. For example, during the conflicts in Ethiopia, Myanmar, and with the Rohingya, Meta’s (Facebook) algorithms were observed to amplify violent content, fueling hatred against the Tigrayans. This emphasis on engaging violent content exacerbates hate speech.
Moreover, newer technologies like generative AIs, capable of independently creating content, disrupt political processes, even in democracies. In the United Kingdom and the United States, “deepfakes” and fabricated audio generated by AI were employed to disrupt political campaigns, undermining the democratic process.
Facial recognition and emotion detection tools, categorized as AI, further exacerbate unequal treatment, particularly when used by governments. Amnesty International criticizes Europe for deploying the “iBorder Control” software at Latvian, Hungarian, and Greek borders, funded by the European Union, which is accused of favoring discrimination.
Roux highlighted that such technologies, labeled as ‘neutral,’ are designed by humans and can be manipulated to target specific groups, posing a threat to vulnerable populations.
Challenges with AI Regulation
Amnesty International underscores the minimal regulation surrounding the widespread use of AI for security issues, which lacks demonstrated effectiveness and necessity. Although the European Parliament adopted the Artificial Intelligence Act on March 24, aiming to establish regulations across all AI domains, Roux believes it falls short, particularly regarding facial identification.
As the Olympic Games approach, France’s decision to legalize AI for surveillance purposes raises concerns for Amnesty International. Despite assurances from the Ministry of the Interior regarding limited usage, the organization fears that such systems may become permanent fixtures in daily life, posing risks to privacy and fundamental rights.
Source: international.la-croix.com
