Amnesty International Calls for Ban on Unlawful Data Scraping Used to Train Generative AI Models

Amnesty International has raised serious concerns over the widespread use of unlawful web-scraping practices by technology companies to collect personal online data for training generative artificial intelligence (AI) systems. In a report released on May 28, 2026, the human rights organisation argued that the large-scale extraction and processing of online information by AI developers violates fundamental privacy rights and other internationally recognised human rights standards.

The report examines the data collection practices underlying generative AI systems developed by major technology companies, including Google, Meta and OpenAI. Amnesty International contends that these systems rely on the mass harvesting of publicly available online content without obtaining the explicit consent of users. According to the organisation, billions of posts, comments, images and other forms of online content have been scraped from the internet and incorporated into AI training datasets, making such systems “unlawful by design and deployment.”

Likhita Banerji, head of Amnesty International’s Algorithmic Accountability Lab, stated that indiscriminate data scraping undermines the right to privacy protected under international human rights instruments and various United Nations resolutions concerning privacy in the digital age. The report further argues that generative AI models reproduce and amplify racial, gender and cultural biases embedded in the datasets on which they are trained, resulting in discriminatory outcomes and reinforcing existing social inequalities.

Beyond privacy concerns, Amnesty highlighted the environmental costs associated with AI development. The organisation noted that the construction and operation of large data centres require significant water consumption and energy resources, while generating substantial amounts of electronic waste containing hazardous materials. It also expressed concern about the extraction of critical minerals used in AI infrastructure, which is often linked to environmentally unsustainable mining practices.

In light of these findings, Amnesty International called upon governments to introduce stronger regulatory frameworks governing AI development and deployment. The organisation urged states to hold technology companies accountable for human rights violations arising from their data collection methods and business practices. It also advocated prohibiting unlawful mass data collection systems used for training standalone generative AI models.

The report forms part of a broader global debate surrounding artificial intelligence regulation, particularly issues relating to surveillance, non-consensual data extraction, profiling, misinformation and algorithmic discrimination. As governments worldwide continue to develop AI governance frameworks, Amnesty has emphasised that innovation must not come at the expense of privacy, human dignity and fundamental rights.