Tech companies hit by $3.5B AI-related fines

AI is advancing faster than the laws meant to govern it, and regulators are starting to push back. Since 2022, courts and data protection authorities around the world have begun issuing fines and settlements against major technology companies for AI-related violations, offering the first real glimpse of where AI development is colliding with existing rules.

This chart of the week examines who has been penalized, which violations regulators are prioritizing, and how enforcement has evolved year over year.

Key insights

• Since 2022, regulators and courts have imposed over $3.5 billion in fines and settlements on seven major technology companies for AI-related violations. Across ten separate cases, Anthropic's $1.5 billion settlement¹ for training AI on pirated books was the largest single penalty, followed closely by Meta's $1.4 billion² for collecting biometric data to train facial recognition without consent. Together, these two cases alone account for 81% of all AI-related fines and settlements in our dataset.
• The vast majority of cases (9 out of 10) centered on a single issue: using data to train AI without proper consent or legal basis. Anthropic ($1.5B), Google ($291M)³, and OpenAI ($17M, later annulled)⁴ were fined for training AI models on copyrighted or personal data without authorization. Meta ($1.4B), Clearview AI ($105M across four cases), and Amazon ($25M)⁵ were fined for collecting biometric data such as facial images and voice recordings to train AI systems. Apple's $250 million settlement⁶ stands apart as the only case where the violation was not about training data, but about overpromising what AI could actually do.
• Not all fines have survived legal challenges. OpenAI's $17 million fine, imposed by Italy's data protection authority in 2024 for training ChatGPT without a legal basis and failing to meet transparency requirements, was later annulled by an Italian court in 2026. The case highlights the legal uncertainty that still surrounds AI regulation: what regulators treat as a violation may not ultimately hold up in court.
• Clearview AI stands out in our analysis as the only company to simply ignore the penalties issued against it by European regulators. In fact, four separate European data protection authorities penalized the company for the exact same violation — scraping billions of facial images to build an AI recognition database. The Dutch DPA⁷ fined Clearview $35 million in 2024, following earlier $23 million fines from the Italian⁸, French⁹, and Greek¹⁰ authorities in 2022, totaling $105 million. To date, Clearview has paid none of them, arguing that because it does not have a physical presence in Europe, it is not subject to European jurisdiction.
• Overall, AI-related penalties began to emerge in 2022, with three fines targeting Clearview AI. In 2023, Amazon was fined for retaining children’s voice recordings used to train Alexa. Enforcement intensified in 2024, when regulators took action in four separate cases involving Google, OpenAI, Meta, and Clearview AI. The trend continued in 2025 with the largest AI-related penalty in our dataset — Anthropic's $1.5 billion settlement. Most recently, in 2026, Apple's $250 million settlement stood apart as the only case about overpromising AI capabilities rather than training data.

Methodology and sources

This study examined AI-related fines and settlements involving major technology companies between 2022 and 2026. We included cases where regulators or courts took formal action over the training, or marketing of AI systems. To qualify, each case had to involve a clearly identified company, a measurable penalty or settlement, and a direct link to AI-related conduct such as unlawful training on copyrighted or personal data, biometric data scraping, or misleading AI claims.

The final dataset contains 10 cases across 7 companies. For each case, we recorded the company, year, penalty amount, and stated reason for the action. We also calculated penalty amounts in both euros and the United States dollars, using a conversion rate of $1 = €0.86. Repeated penalties against the same company were counted separately when they came from different authorities. Annulled penalties were also included, but marked accordingly, to reflect the full enforcement landscape.

References: