AI experts discuss gaps in the EU AI Act
The EU is right to be proposing an Artificial Intelligence Act but its current approach could risk creating more legal gaps AI experts have warned.
Malta’s IT Law Association (MITLA) recently held a webinar about AI, featuring a panel of distinguished technology law experts including Odia Kagan from the USA, Peter Hense from Germany and Antonio Ghio from Malta.
Hense, partner at Spirit Legal LLP, Leipzig, Germany said: "The European AI Act has evolved into a 'bureaucratic monster' overwhelmed by lobbying. What started as a lean regulation is now a 265-page
document with significant gaps for evading regulation. The real essence of this regulation lies in the standards it refers to, which are not fully developed yet.”
Hense suggested that the next five years would witness significant changes driven by data protection laws and a renewed focus on privacy in the US and Europe, impacting the AI ecosystem and big tech profits.
He emphasised the need for building more transparent AI models, advocating for integrating transparency into the fabric of AI development rather than creating separate tools for transparency.
Kagan, partner and chair of GDPR Compliance and International Privacy at Fox Rothschild LLP, Philadelphia, USA, said: "In the US, our approach has
been to recognise smart AI will essentially perform functions that humans do. The US approach is: if it’s illegal when humans do it, it's illegal with AI. This practical approach shapes our current AI regulation landscape.”
Kagan also discussed how the hype around the success and limitations of language models like ChatGPT could be distracting us from the bigger innovations that could have more existential consequences in the world.
She expressed hope that 2024 would be a year of increased awareness about AI, preparing for more sophisticated AI in the future. She mentioned the potential of the current hype around AI to serve as a catalyst for raising global awareness and laying the groundwork for
future regulations and ethical considerations in AI deployment.
Ghio, partner at Fenech & Fenech Advocates, Malta said: "For Malta, and by extension the EU, the challenge is not just adopting regulations but ensuring that these regulations are aligned with technologyneutral principles. The risk is that by regulating too early, we might inadvertently hamper innovation initiatives.”
Ghio highlighted the ethical dimensions of AI, stressing the importance of having the right principles in place to guide AI development and deployment.
He pointed out the differences between AI regulation and data protection laws, noting that while data protection is about fundamental rights of every individual, AI regulation is about
managing tools deployed by specific companies. He emphasised the need for a balanced approach which maintains ethical and legal standards while allowing for safe innovation.
The discussion was chaired by Christian Peregin and was followed by the MITLA AGM, which elected the following candidates to the executive committee: Antonio Ghio, Gege Gatt, Deo Falzon, Sarah Cannataci, Keith Cutajar, Kenneth Terribile, Sharon Xuereb, Mark Zammit and Veronica Campbell.