Global AI rules diverge as UK unveils justice sector action plan
As artificial intelligence regulation accelerates across the globe, the coming months will see the implementation of the European Union's new rules for general-purpose AI models. Thought leaders from technology and justice sectors are weighing in, highlighting both the promise and complexity of today's rapidly diverging regulatory landscape.
Helena Nimmo, Chief Information Officer at IFS, a global leader in industrial AI, framed the current regulatory moment as a "trifurcation of policy" between the European Union, United States, and China. "Each approach has its pros and cons. The US model is focusing heavily on innovation, the EU model emphasises safety and security, and the Chinese model focuses on state-controlled AI leadership. Personally, I think it is too early to call on the success of any of the approaches. AI is in its infancy, and I am sure we will see a v2.0 of the frameworks emerging fairly rapidly," Nimmo said.
Experts note that while the US has released a Trump-era AI Action Plan designed to boost innovation, it contrasts sharply with the EU's focus on risk and accountability, and China's highly centralised control. Bob De Caux, Chief AI Officer at IFS, agreed on the distinct impacts each framework would have for industry. "Broadly, the US AI plan will have more impact for companies harnessing the upstream infrastructure needed to build and iterate on foundational AI models, while the EU AI act will affect companies deploying AI applications downstream, particularly in high-risk industries," De Caux explained. "As the large AI foundation model providers are increasingly reaching out downstream, they won't be able to avoid being impacted by the EU AI Act when they are delivering to customers within Europe. However, being based in the US might still give them benefits in being able to iterate and test products in a more deregulated market, while they build a compliance offering that allows them to deliver by region."
This regulatory divergence is expected to have significant implications for tech firms operating internationally. Providers of large-scale AI foundation models may find themselves navigating multiple, and at times conflicting, compliance regimes, needing to adjust products and services for different regulatory geographies. The result may be a wave of rapid innovation in compliance solutions and cross-border regulatory technology.
Meanwhile, in the United Kingdom, policymakers are grappling with the deployment of AI in sensitive public sector domains. The Ministry of Justice has unveiled its AI Action Plan, mapping out a vision for how artificial intelligence could reshape criminal courts and wider justice services. Katherine Beadle, Client Partner for Justice at NTT Data UK&I, described the plan as a "promising step toward modernising the wider justice system," but emphasised the importance of ensuring changes "genuinely benefit both the public, and the people who keep the system running – rather than replacing them."
Beadle elaborated, "It is both ambitious and action-oriented which in turn should deliver more efficient services where they're needed most, from case management to rehabilitation and community safety. Crucially, it recognises that technology and AI isn't about replacing workforces, but empowering those central to keeping the system moving." She highlighted that tools such as automated transcription and smart document search could reduce administrative burden, improving workflow for probation officers, court staff, and others on the frontline.
Still, Beadle cautioned that swift justice must not come at the expense of fairness. "That means putting safety and fairness first by working within the law to protect individual rights and maintain public trust. That requires careful oversight and a clear framework for accountability, especially when decisions impact liberty." The 'Scan, Pilot, Scale' model adopted in the plan, she said, addresses these concerns by allowing new systems to be tested in controlled conditions prior to broader rollout, balancing effectiveness with responsibility.
Looking ahead, partnerships between government, regulators, and technology providers will be key to moving beyond incremental change towards lasting transformation. "AI is a powerful tool when implemented the right way, and it will be essential to delivering lasting transformation and accelerating justice for victims," Beadle said.
The coming year will test the mettle of these new frameworks, both in industry and in the public sector. As Nimmo concludes, with AI still in its infancy, the global community may see the emergence of second-generation governance frameworks sooner than anticipated, as governments and businesses alike grapple with the challenge of harnessing AI's potential while safeguarding the values and rights at the heart of society.
