To regulate or not to regulate, that is the question

Over recent years, artificial intelligence (“AI”) has become an increasingly prominent part of our everyday lives. With this increased prevalence comes a debate around whether the technology should be subject to regulation or not. With the initial phases of the EU AI Act (the “Act”) implementation having taken place earlier this year, we take a look at the arguments for and against the UK adopting a similar approach and outline its current plans with regard to AI regulation.

The EU AI Act

The EU AI Act came into force on 1 August 2024 and is the first comprehensive regulation on AI anywhere in the world. On 2 February 2025, the first two chapters of the EU AI Act became applicable, including an outright ban on specific prohibited practices and an AI literacy requirement.

The EU AI Act takes a risk-based approach, classifying AI systems based on four risk categories ranging from unacceptable to minimal or no risk. These categorisations subsequently inform which sections of the EU AI Act are applicable to a given business and the obligations that must be met. A distinction is also drawn between six different roles of organisations within the AI supply chain including providers, deployers and distributors.

Much like the General Data Protection Regulation, the EU AI Act has extraterritorial effect, meaning that it does not only apply to EU based businesses but also those established outside of the EU where the output of their AI system is used in the EU or affects persons located in the EU.

The UK approach

On 3 August 2023, the (conservative) government at the time published a white paper entitled “A pro-innovation approach to AI regulation”, outlining a slightly different approach to that found in the EU. Instead of implementing statutory rules, it opted for a more flexible approach to regulation, identifying a framework, focussing on 5 principles aimed at guiding and informing responsible AI use and development across all sectors. Existing regulators (such as the Information Commissioners Office) will be responsible for adapting these principles to specific contexts, such as data protection, and enforcing them where necessary.

However, this approach is not favoured by all with some, notably Lord Holmes of Richmond, advocating for the regulation of AI. During the aforementioned conservative government, he proposed the Artificial Intelligence (Regulation) Bill, a private members’ bill which sought to establish an AI Authority to address AI regulation in the UK. As a result of the UK general election on 4 July 2024, the bill was dropped, although it was reintroduced on 4 March 2025 and appears to have a more positive outlook this time around but the intention is that it will have narrower application, focussing on developers of frontier AI systems.

This has left many questioning why the UK is not regulating AI in a similar way to the EU and what the potential impact of this might be.

 Benefits and drawbacks of comprehensive AI legislation

AI as a technology is currently still in its infancy with new applications and use cases emerging frequently. This is achieved through constant innovation, a process that requires investment to be effective and this forms the foundation of many arguments against the adoption of a UK AI Act.

By imposing rigid rules at such an early stage, investment and, in turn, innovation are likely to be slowed, potentially resulting in the UK losing its current status as a global AI leader. Additionally, any rules that are adopted are likely to become obsolete or unduly restrictive within a short timeframe, potentially focussing on the wrong use cases or leaving scope loopholes that can be exploited. All of these risks are, to some extent, mitigated by the adoption of an informal code of practice which can adapt to future risks and technologies as and when they arrive.

On the opposite side of the fence, some argue that not providing clear rules in the form of legislation could fail to mitigate the risk of harm to individuals in critical sectors such as healthcare, finance and criminal justice. Concerns include algorithmic bias and a lack of transparency around decision making practices, issues that have arisen in relation to AI for many years. By placing the rules on a statutory footing, organisations will have a greater degree of certainty and the UK’s valuable start-up ecosystem will be better able to compete with the tech giants that dominate much of the industry at present.

Conclusions

Whilst there are convincing arguments in favour of both approaches, one approach could be to adopt a middle ground. As outlined above, the pro-innovation approach currently taken by the UK provides much needed flexibility but the government itself also needs to remain flexible, considering whether voluntary compliance is sufficient and, if not, reviewing its approach. Additionally, providing sector-specific regulators with the degree of discretion currently suggested could lead to a lack of consistency that could be reduced by the creation of a central AI oversight  body.

Irrespective of how the UK proceeds, many UK organisations will be subject to the EU AI Act and, as a result, will need to achieve compliance. With the first stage of implementation now complete, considering your compliance obligations and how they can be met at an early stage is crucial. However, this is only one piece of the puzzle.

In addition to the EU AI Act, organisations need to consider how existing UK laws apply to their use of AI, implement robust governance procedures early in the adoption lifecycle and ensure knowledge of, and compliance with, UK regulator guidance and recommendations. Cripps has developed a six-step process to guide you through the AI governance journey, ensuring that you have the documentation in place to support the effective and ethical use of AI. It uses a combination of guidance (to help you understand what’s necessary, advisable, and optional), policies (for you to customise and implement) and records (for you to complete and maintain).

How we can help

If you would like to speak to one of our AI experts or make an enquiry about the AI toolkit, please use the contact form available at AI governance toolkit.