Leaked EU AI Act is ‘a wake-up call for organisations’

A leak of the EU’s AI Act removes doubts about its adoption, but critics fear that the plans will inhibit innovation.

All 892 pages of the final version were published on Monday by Euractiv tech editor Luca Bertuzzi, while a consolidated version that runs a mere 258 pages was shared by European Parliament policy advisor Laura Caroli. Only linguistic or technical details can now be made to the full regulation.

EU ambassadors are due to adopt the act on February 2. If all goes to plan, the final regulation will be implemented before the European Parliament elections in June.

That leaves businesses and organisations with little time to prepare for the new requirements. Experts warn that they should start planning immediately.

“The EU AI Act will become law shortly,” Enza Iannopollo, principal analyst at business advisory firm Forrester, told TNW.

“More importantly, the leak is a wake-up call for organisations to start working on it, now. Future changes to the text will not affect the substance of the act.”

Some provisions will become enforceable earlier than others. Iannopollo recommends taking a close look at the act’s enforcement timeline.

“It will take time to get all of this going,” Iannopollo said. “Organisations must use this time to put together an AI compliance task force, build a roadmap for compliance on time, and prioritise work on the most complex aspects of the EU AI Act.”

Global repercussions

Organisations outside the European Union also need to get a move on. As the regulation has extraterritorial applications and could cover any sector, it’s not only EU tech firms that need to follow the rules.

Because of this global scale and the act’s diverse requirements, the demand for AI governance programs and professionals is set to boom.

A prominent example was exposed today. According to another document leaked to Euractiv, the EU plans to establish an AI Office to support the new regulation.

“As AI technologies become more ubiquitous, and the governance ecosystem more mature, we can expect many more organisations to carve out dedicated AI governance and personnel,” Joe Jones, Director of Research and Insights at the International Association of Privacy Professionals, told TNW.

AI risks and rewards

All these additional requirements also pose a threat to the future of EU tech.

If the regulation creates excessive burdens and risk-averse environments, it could restrict the benefits of AI.

“Unsurprisingly, the act contains many details, and let’s hope this does not turn out to be an ‘EU innovation stopper’,” said Dr Andreas Splittgerber, partner at global law firm Reed Smith.

It’s a concern that’s been echoed by leading EU politicians. In December, French President Emmanuel Macron warned that the AI Act imperils European tech companies.

“We can decide to regulate much faster and much stronger than our major competitors,” he said. “But we will regulate things that we will no longer produce or invent. This is never a good idea.”