What regulators have actually collected from Big Tech over AI

Enforcement over AI is no longer prospective. Research published in 2026 tallies more than $3.5 billion in AI-related fines and settlements imposed on seven major technology companies since 2022.

The shape of that total is worth reading carefully, because it is not what current AI legislation is mostly about. The largest single item is Anthropic’s $1.5 billion settlement in 2025. Enforcement began in 2022 with actions against Clearview AI over facial recognition, and in 2023 Amazon agreed to a $25 million civil penalty with the FTC and the Department of Justice over retention of children’s voice recordings and location data from Alexa. 2024 brought actions involving Google, OpenAI, Meta and Clearview AI; OpenAI’s $17 million fine from Italy’s data protection authority, for training ChatGPT without a legal basis and for transparency failures, was later annulled by an Italian court in 2026.

One 2026 case stands apart: Apple’s $250 million settlement concerned overpromising AI capabilities rather than training data — an advertising and consumer-protection theory rather than a privacy one.

The pattern underneath is that almost all of this money was collected under laws that predate the AI statutes now taking effect: data protection, children’s privacy, biometric consent, consumer protection. The AI-specific regimes are only now acquiring enforcement powers. For a company estimating exposure today, the operative rules are still the old ones.