EU targets smooth transition to AI Act with voluntary pact

The AI Act entered into force at the beginning of August, and places legal requirements and obligations on companies that want to develop and deploy artificial intelligence. However, some provisions within the AI Act will take up to three years to enter into force, and so in an effort to ensure a smooth transition, the EU has launched the AI Pact.

This is a strictly voluntary initiative designed to help the private sector get its collective ducks in a row before the rules come into force.

There are two pillars to the Pact. The first aims to create a collaborative community that can create and share best practices and internal policies when it comes to preparing for life under the AI Act. The second pillar aims to promote and foster the early implementation of AI Act provisions before they enter into force.

It's all being done in the name of responsible AI development and governance.

The full list of initial signatories can be found here. The telcos currently on there are Deutsche Telekom, Orange, Telefónica, Telenor, TIM and Vodafone. Noteworthy network vendors, device OEMs and tech giants include (deep breath) Amazon, Cisco, Google, HPE, IBM, Lenovo, Microsoft, Motorola Solutions, NEC, Nokia, OpenAI, Qualcomm and Samsung.

Nokia is particularly keen to publicise its participation, welcoming an opportunity to contribute to responsible AI development.

"Nokia's AI Pact pledges underscore our ongoing commitment to the responsible, innovative and business-oriented implementation of artificial intelligence in our AI-powered products and services," said Ingrid Viitanen, general counsel, Nokia Strategy & Technology.

"Nokia has set up an AI Governance Framework programme to strengthen our internal AI-related processes and build trust with stakeholders, including customers, suppliers and authorities," she said. "In parallel, we continue to contribute actively to building industry standards reflecting the AI Act's requirements."

"We look forward to sharing our learnings collaboratively and transparently with our industry peers and with the EU AI Office."

Meanwhile, conspicuous by their absence from the list at this point in time are Apple, Ericsson, and Meta. Those last two in particular stick out like sore thumbs, since they were among a list of signatories on an open letter criticising the EU's approach to AI regulation.

This letter accuses the EU of "inconsistent regulatory decision making," creating uncertainty about what data can be used to train AI models, and what data threatens to put companies on the wrong side of the EU's data protection rules. This lack of clarity risks delaying AI deployments within the EU, putting the bloc at a competitive disadvantage to its global peers.

As much as the EU can be accused of being overbearing, when it comes to overseeing AI, it is caught between a rock and a hard place.

Having been told for years that AI has the power to revolutionise every sector of the economy and even tear up the very fabric of society, it is understandable that Brussels doesn't want to be accused of being asleep at the wheel.

However, by trying to get out in front of the technology and establish a regulatory framework, it can be accused of inadvertently stifling AI development with too many rules and restrictions.

There's probably a happy medium in there somewhere, the EU could do with hurrying up and finding it.