Generative AI "Fundamentally Incompatible" with International Human Rights Law, Amnesty Calls for Ban

Amnesty Denounces Human Rights Issues in

Generative AI On May 31, 2026, international human rights organization Amnesty International published a report titled “Unlawful by Design: Exposing the Human Rights Costs of Generative AI,” which thoroughly analyzes the impact of generative AI systems on human rights. The report concludes that standalone generative AI systems, which rely on illegal web scraping, clash with international human rights law (IHRL) and related standards throughout their design, development, and operation. Amnesty asserts that these systems are “fundamentally incompatible” with international human rights law and has called for their prohibition.

Why Are These Systems Deemed “Illegal”?

At the core of the report is the legality of the data collection methods relied upon by generative AI systems. Many major generative AI models currently build their training data by automatically collecting vast amounts of text and images published on the internet through web scraping, often without the consent of copyright holders. Amnesty highlights that such data collection methods are “inherently dependent on large-scale privacy violations.” While web scraping is not a new technology, its scale and purpose in the context of generative AI are fundamentally different. Text data from blogs, forum posts, social media comments, academic papers, and news articles—essentially all publicly available internet content—can potentially be exploited as training data without individual consent. Amnesty has determined that this practice violates privacy rights protected under international human rights law.

Three Key Dimensions of Human Rights

Violations The report focuses on three key aspects of human rights violations caused by generative AI systems. 1. Privacy Rights Violations: As previously mentioned, Amnesty considers the unauthorized use of personal content for AI training to erode international data protection standards. 2. Promotion of Discrimination: Generative AI models often incorporate biases and discriminatory expressions present in their training data, which are collected from the internet. Consequently, the output of these AI systems risks reflecting biases related to race, gender, religion, and other factors, potentially amplifying institutional discrimination. 3. Threats to Freedom of Expression and Thought: The content generated by AI could distort informational environments, unfairly influencing individual thought processes and the expression of opinions. In particular, the ease with which misinformation and deepfakes can be generated poses a serious threat to the foundations of democratic discourse.

The Cost of Industry “Efficiency”

The report sharply questions the hidden costs behind the promised “sophisticated automation and efficiency” of generative AI. The tech industry has promoted generative AI as a tool for enhancing productivity and streamlining operations. However, according to Amnesty’s analysis, these benefits are built on data collection and model training practices that violate privacy rights, promote discrimination, and threaten freedom of expression. Amnesty takes issue with the structural problem of sacrificing human rights in the name of technological innovation. This reflects a “move fast and apologize later” approach typical of Silicon Valley, which has led to significant friction with international human rights norms.

The Significance of Calling for a “Ban”

While Amnesty has published numerous reports on human rights, the explicit call for the “ban” of certain technologies is particularly noteworthy. The organization has concluded that standalone generative AI systems that rely on illegal web scraping are fundamentally incompatible with international human rights law and has explicitly called for their prohibition. Rather than advocating for limitations on parts of generative AI, Amnesty’s report conveys a strong message that the prevailing development methods themselves are unacceptable under human rights law. Globally, various nations, including the EU, are exploring frameworks for AI regulation, such as the AI Act. However, Amnesty’s report argues that existing regulatory approaches are insufficient and calls for a more decisive stance: banning the use of technologies that lead to human rights violations.

Impact on the AI Industry and Future Debate

The report is likely to have significant consequences for the AI industry. Major AI companies are already facing lawsuits worldwide over the legality of their data collection practices. Amnesty’s declaration that these systems are “fundamentally incompatible” with human rights law adds further legal and public pressure. On the other hand, some generative AI developers argue that using publicly available data constitutes fair use and claim that improved data collection methods could align AI development with human rights. The extent to which Amnesty’s report challenges these industry claims will be a focal point in future debates. The relationship between technology and human rights is one of the most pressing issues in the age of generative AI. Amnesty’s report has the potential to significantly advance discussions on this critical topic.