The European Commission has revealed the list of the first 100+ signatories to the AI Agreement. The agreement focuses on forcing companies to make “voluntary commitments” regarding how they approach and implement artificial intelligence.
The bloc’s legally binding Risk-Based AI Rulebook (AI Act) came into force last month, but it will be several years before all compliance deadlines are implemented. This has created a vacuum of non-compliance, which the EU is trying to fill with the AI Agreement.
The initiative aims to increase engagement and drive commitment so businesses can stay ahead of the curve by taking steps to implement the law’s requirements before the deadline. The agreement also focuses on facilitating information sharing so that the signatories can collaborate with each other to address the emerging requirements of the bloc’s AI rulebook and proactively develop best practices.
There are also three “core actions” that signatories to the agreement are expected to commit to (at least).
Adopt an AI governance strategy to accelerate the adoption of AI in your organization and work toward compliance with future AI laws. Identify and map AI systems that may be classified as high risk under the AI Act. Promote AI awareness and literacy among staff and ensure ethical and responsible AI development.
Additionally, there is a long list of potential pledges (available here in PDF format), which, according to the European Commission, have been drafted by the AI Directorate, the body that oversees AI law, and which, according to the European Commission, are part of the AI Agreement’s ‘relevant The selection was made after receiving feedback from “stakeholders.” network. The resulting list of pledges essentially allows signers to choose and combine which pledges work for them.
Examples include “designing AI systems that are intended to interact directly with individuals, so that individuals are notified on demand that they are interacting with the AI system,” and “developing deep information.” Includes a pledge to clearly and distinctively label the images, audio, video, and other AI-generated content that it consists of. fake. ”
This long list could foster compliance-promoting competition among signatories to see who offers the most services when it comes to AI safety.
Agreement to promote rapid compliance with AI laws
The AI Pact initiative was unveiled by then Internal Market Commissioner Thierry Breton in May 2023, and Google agreed at the time to help regulators work on the initiative. More than a year later, the EU has collected more signatures, but some prominent names are missing from the list.
For example, Apple is not publicly traded and neither is Meta. The ad tech giant told Reuters on Tuesday it would not immediately join the effort, saying it wanted to focus on compliance efforts under the AI Act itself.
The penalties for violating the EU’s legally binding AI rulebook are severe and can reach up to 7% of global annual revenue for violating the prohibited use of AI. Up to 3% for non-compliance with other AI law obligations. Up to 1.5% for providing false information to regulators.
So, if Meta takes a wrong step when it comes to actual AI rules, it could face billions of dollars in fines. Breaking on commitments could simply lead to dressing down in public, which may be why it has so far ignored the agreement.
Mistral, a major French language modeling company, is also not on the list. The company has been one of the most vocal critics of the AI Act, so it’s not too surprising that it hasn’t signed the voluntary pledge.
Meanwhile, Germany’s Aleph Alpha, another major European language model maker, also signed an agreement. However, the company recently announced that it is pivoting to providing B2B support for generative AI tools. Policy priorities may also be reprioritized in light of evolving business models.
Others on the list include Amazon, Microsoft, OpenAI, Palantir, Samsung, SAP, Salesforce, Snap, Airbus, Porsche, Lenovo, and Qualcomm.
Meanwhile, Anthropic, Nvidia, and Spotify are nowhere to be seen. The absence of the first two companies is particularly notable given the importance of AI development.
Spotify’s absence is notable as the European company last week signed an open letter sponsored by Meta, which lobbies against regulations that could curb the spread of generative AI.
A complete list of the EU’s initial AI Pact signatures can be found here.
Various types of companies have registered, including large European telecom companies, consulting companies, software companies, banking/payment companies, multinationals, small and medium-sized enterprises, and consumer platforms. Given how far and fast generative AI technology is spreading, it’s clear that over 100 names are just the tip of the iceberg.
And since these are purely voluntary pledges, signing an AI agreement may mean little more than a bid to gain reputational influence. Signatories will also be asked to report on their progress 12 months after announcing their own pledge, opening up another opportunity for promotion.
