UK Government Seeks Expanded Powers to Regulate AI Under Online Safety Act

The United Kingdom government is moving to grant ministers broad new authority to amend the Online Safety Act (OSA) through provisions inserted into two separate pieces of legislation. These proposed changes, which critics warn could bypass standard parliamentary scrutiny, are designed to allow the government to address rapid developments in artificial intelligence, including concerns surrounding AI-generated content.

Proposed Amendments and Legislative Strategy

The government has introduced amendments to the Crime and Policing Bill and the Children’s Wellbeing and Schools Bill that would grant ministers the power to rewrite significant portions of the Online Safety Act. By utilizing “Henry VIII clauses,” the government could theoretically bypass full parliamentary debate, limiting the legislature to a simple yes-or-no vote on new regulatory rules.

Lorna Woods, a law professor at Essex University and legal advisor to the Online Safety Network, has highlighted the scale of these potential changes. According to Woods, these powers could effectively introduce a substantial volume of new regulatory requirements, allowing ministers to add unforeseen rules without the typical legislative process.

The push for these powers follows admissions by Ofcom, the UK’s communications regulator, that the current framework of the Online Safety Act did not provide sufficient authority to address specific harms, such as non-consensual sexualized imagery generated by AI chatbots like Grok. The government has stated its intention to close these legal loopholes to ensure all AI chatbot providers adhere to the illegal content duties mandated by the OSA.

Debate Over Democratic Scrutiny

The move has sparked a wider conversation about the balance between executive speed and democratic accountability. While proponents of the amendments argue that the rapid evolution of online threats—particularly those involving AI—necessitates a more agile regulatory response, critics express concern that the shortcut approach undermines the legitimacy of the regime.

Javier Ruiz Diaz, technology and human rights lead at Amnesty International UK, described the existing act as a complex, “Frankenstein” system. He warned that bypassing standard procedures to add further amendments could create gaps that tech companies might exploit, ultimately weakening the enforceability of online safety standards.

experts like Elena Abrusci, a senior lecturer in law at Brunel University, have noted that while some ministerial power to update secondary legislation is not unprecedented, the specific focus on AI-generated content represents a significant expansion of executive control. There are fears that limiting parliamentary debate will make it easier for large technology firms to lobby ministers directly, rather than engaging with a transparent and public legislative process.

The Future of Online Safety Regulation

The Online Safety Act, which became law in 2023, was the result of years of deliberation. As the government attempts to integrate AI regulation into this existing structure, it faces the challenge of maintaining public trust. Owen Bennett, former head of international online safety at Ofcom, emphasized that the success of the act has historically relied on the perception that it reflects public will.

As the government navigates these legislative changes, several key concerns remain at the forefront of the debate:

• Regulatory Scope: Experts worry that amendments tied to specific bills, such as the Crime and Policing Bill, may result in “partial solutions” that fail to address the broader societal risks of AI, such as mental health impacts and addiction.
• Institutional Oversight: There is an ongoing call for regulators like Ofcom and the Equality and Human Rights Commission to maintain rigorous oversight to ensure the government does not overstep its authority.
• International Impact: The UK’s approach to AI regulation is being watched globally, with concerns that a perceived shift toward executive control over tech regulation could set a challenging precedent for democratic digital governance.

For now, the government maintains that the necessity to act quickly against emerging online harms justifies the proposed legislative path. However, the tension between the need for speed and the requirement for robust parliamentary oversight remains a defining feature of the UK’s evolving digital policy landscape.