The European Commission’s current draft AI Act can promote AI innovation across Europe if nuanced amendments are made – Xperi issues call to action
As the European Commission’s draft regulatory framework on artificial intelligence (AI) Act continues through the EU legislative process – where it is currently under review by the European Parliament – the draft AI Act has already been subject to scrutiny by various organisations across Europe and beyond and, most recently, subject to suggested modifications from the Slovenian and French Presidencies. John Patterson, VP and Associate General Counsel EMEA at Xperi, is proposing further amendments to the draft AI Act, which, if adopted, he argues will ultimately promote AI innovation across Europe.
Patterson’s amendments, include:
- Adding subtlety into the classification of high-risk AI systems to ensure that the proposed rules on ‘high-risk’ AI only applies to AI systems that pose an actual risk of harm to health and safety or a risk to fundamental rights
- Applying the same EUR 500,000 penalties for non-compliance to commercial and government organisations alike to create a fair playing ground and promote balanced AI innovation across Europe
“While we welcome a regulatory framework on AI, and commend the European Commission for their brave and comprehensive first draft, as well as the Slovakian and French Presidency changes, it is critical to the business and citizens of Europe that the European Commission strikes the right balance. Subtle amendments to the draft AI Act are required, otherwise we could see AI Innovation across Europe be severely impacted.
“The current draft AI Act, as modified by the Slovakian and French Presidencies, still raises concern around the definitions regarding high-risk AI, and imposes hefty and unnecessary fines on companies, not governments, which could cause European AI innovation, in various areas which may not be ‘high-risk,’ to lag behind the rest of the world. If changes are not made, it could be particularly detrimental to SMEs, who can least afford unnecessary costs of non-compliance.
“If the European Commission adopts the wording we’re proposing, we believe this would preserve the goal of creating harmonised EU wide AI legislation that would heavily regulate high-risk/bad AI yet allow SMEs and other organisations within Europe to develop areas which are not actually ‘high-risk’ – either because of their nature or because mitigations put in place. Put simply, we need to act now, before it’s too late.”
For an in-depth overview of John Patterson’s suggested amendments, please visit the Xperi website:
- AI Act