Over 100 companies have signed the EU’s Artificial Intelligence (AI) pact, including Google and Microsoft, but Meta has not yet joined.
The European Commission invited companies last year to make voluntary commitments for the safe and reliable use of AI, anticipating the full implementation of new European legislation by 2027. This law has been in effect since August, but its provisions are being gradually applied over the coming years.
More than a hundred companies have already signed the pact, agreeing to expedite the application of the legislation’s principles. These include major companies like Google, Amazon, and Microsoft, as well as small and medium-sized enterprises from various sectors (automotive, aviation, banking, etc.).
Signatories commit to actions such as mapping high-risk AI systems and educating their staff. Over half of the signatories have pledged further steps, including human oversight of AI, risk mitigation, and clear labelling of certain AI-generated content, such as ultra-realistic video fakes ("deepfakes").
Meta has not yet signed the pact. A spokesperson stated that the company is currently focusing on complying with the AI legislation. Meta’s distinctive AI model, Llama, features open-source functionalities that users can reuse with minimal developer control, which could complicate risk assessments.