Over 100 EU Firms Commit to Implementing Key AI Act Provisions

More than 100 tech companies on Wednesday made voluntary commitments to conduct trustworthy and safe development of artificial intelligence in the European Union, with a few notable exceptions.

See Also: OnDemand | Fireside Chat: Staying Secure and Compliant Alongside AI Innovation

The AI Pact, signed by OpenAI, Microsoft and Amazon, is a voluntary alliance that aims to help businesses in the European Union prepare for the AI Act as the law’s requirements stagger into enforcement over the next two years.

The European Commission reportedly expects about 700 companies to join the voluntary pact.

Tech giants Meta, Apple, Nvidia and Mistral had not signed the pact at the time of its release.

The AI Act, which became law on Aug. 1, looks to regulate models based on the risks they pose. Until the full effect of the law kicks in, the European Commission expects the voluntary AI Pact to help companies implement the key obligations of the world’s most comprehensive artificial intelligence legislation.

As part of the pact, the participating companies should collaborate to share ideas on best practices for compliance, identify systems likely to be categorized as high-risk and promote AI literacy.

More than half the signatories also committed to ensuring human oversight, mitigating risks and transparently labeling certain types of AI-generated content, such as deepfakes, the commission said.

The pact has representation from companies in the telecom, software, banking and payments sectors, along with consulting firms, software players, multinationals and small and medium businesses.

Mistral’s absence is not surprising, since it has been one of the biggest critics of the act. The absence of Apple and Nvidia is notable because both companies have supported AI regulations and are among the signatories in the United States’ bid to develop the technology in a safe and responsible manner. Meta told Reuters that it will not immediately be part of the pact and instead will focus on complying with the larger AI Act.

The penalties for not complying with the bloc’s legislation are stringent – up to 7% of a company’s global annual revenue if it violates the rules on banned uses of AI and up to 3% for noncompliance with other AI Act obligations. Sharing incorrect information with regulators can cost the violating company 1.5% of its annual revenue.

The pledges taken as part of the AI Pact are voluntary, and not following through on the promises has no consequences. The signatories are expected to report their progress in a year.