The European Union's draft compromise on the AI Act reveals that tech businesses and startups working on powerful AI models, such as GPT-4 in Europe, will face additional legal scrutiny. This move makes Europe one of the first regulatory areas globally to oversee such advanced AI models and the proposed regulations primarily affect American companies, as European players like Mistral, utilising open-source models, are exempt. The compromise introduces a two-tiered approach categorising AI models as general-purpose with or without "systemic risk," the latter being models with high-impact capabilities. In light of this, Claire Trachet, tech expert and CEO/Founder of business advisory firm Trachet, discusses the significance of the EU’s new guardrails and what this means for the UK.
Recently, the UK's cyber security agency sounded the alarm on the escalating threat posed by artificial intelligence to the integrity of the next general election, which must happen by January 2025. The National Cyber Security Centre (NCSC) also warned that other major elections in Western democracies, such as the United States, are also under threat. The NCSC, part of GCHQ, emphasised the urgent need for heightened vigilance and a comprehensive review of the electoral process in the face of evolving AI capabilities, a warning which clearly the EU has taken seriously.
The EU's AI Act negotiations, scheduled for completion this week, have faced contention, particularly regarding foundation models like GPT-4. Originally designed to impose risk-based requirements on AI applications, the emergence of AI chatbots prompted European Parliament members to propose additional rules for foundation models, irrespective of their application risk. This shift raised concerns among startups and industry groups, leading France and Germany to temporarily block talks to safeguard their startups, including Mistral and Aleph Alpha. The final deal is expected to be announced soon, with the aim of preventing legislative delays until after the upcoming European elections.
Claire Trachet, CEO/Founder of business advisory, Trachet commented: “Cyber security stands as the guardian amidst growing threats in the face of ever-advancing AI. It is encouraging to see the EU taking proactive measures amidst continuously evolving technology, once again proving its potential to be a global leader in the AI space.
“The importance of cyber security extends far beyond a consideration, it is imperative in an evolving digital society. As we navigate an increasingly complex and interconnected world, the call to fortify our cyber defences resonates louder than ever. it is vital that the UK government follows the example set by the EU and channels a greater level of investment into cyber security to protect both UK businesses and democratic processes from potential interference.”