U.K.’s AI Safety Ambitions Clash With Lack Of Domestic Regulation, Report Reveals

In recent weeks, the U.K. government has been promoting its international role in AI safety while neglecting to pass new domestic legislation for regulating AI applications.

The government’s policy paper on the topic claims to be “pro-innovation” but lacks substantive rules to guard against risks and harms.

These contradictions are highlighted in a new report by the Ada Lovelace Institute, which provides recommendations for strengthening the U.K.‘s approach to AI regulation.

The report emphasizes the need for a comprehensive definition of AI safety, considering the wide range of harms that can arise as AI systems become more embedded in society.

Also Read

• Using AI to Tackle Wildfires: Scientists Embrace Collaborative Solutions
• Stability AI CEO Predicts Artificial Intelligence To Become The “Biggest Bubble Of All Time”

It focuses on regulating the real-world harms caused by AI systems today, rather than speculative future risks. The Institute suggests that the U.K. needs effective domestic regulation to establish itself as an “AI superpower” and create a robust AI economy.

The report outlines 18 recommendations for improving the U.K.’s current approach to AI regulation.

It criticizes the government’s reliance on existing regulators to interpret and apply broad principles without new legal powers or additional resources. In contrast, the European Union is actively developing a risk-based framework to regulate AI.

The U.K.’s approach raises concerns about regulatory inconsistency and uncertainty for AI developers. The lack of clarity on which existing rules apply to AI applications could lead to confusion and increased costs.

The report also highlights gaps in the U.K.’s regulatory landscape, particularly in sectors such as recruitment, education, policing, and unregulated parts of the private sector.

Furthermore, the U.K.’s ambition to become an AI safety hub is undermined by efforts to weaken data protection regulations.

The proposed deregulatory reform of the national data protection framework, known as the Data Protection and Digital Information Bill (No. 2), reduces the level of protection for individuals subjected to automated decisions.

The Institute warns that this undermines the government’s regulatory proposals for AI.

Contradictions in U.K. Government’s Approach to AI Safety Highlighted by Ada Lovelace Institute

The Ada Lovelace Institute recommends that the government reconsiders elements of the data protection reform bill that may undermine AI safety.

It calls for a statutory duty for regulators to consider the AI principles, increased funding and resources for regulators, the exploration of a common set of powers, and the establishment of an AI ombudsperson.

The report also suggests mandatory reporting requirements for foundational model developers and government investment in pilot projects.

The report concludes that the U.K.’s credibility in AI regulation relies on its ability to deliver a world-leading regulatory regime domestically.

International coordination efforts are welcomed but not sufficient on their own. To be taken seriously and achieve its global ambitions, the U.K. must strengthen its domestic regulatory proposals for AI.