AI and Human Rights

On 14 September 2026, the UK Parliament's Joint Committee on Human Rights (JCHR) published a long-awaited report on AI and human rights, warning that the UK’s current legal and regulatory framework does not offer adequate protection against the human rights risks that AI systems pose. According to Committee Chair Alex Sobel MP, “nowhere in the world, including the UK, has a current legislative and regulatory approach to AI that is fit for purpose”.  

Following an inquiry launched in July 2025, the JCHR’s report recommends the introduction of a dedicated AI Bill and the establishment of a single, independent AI regulator, with enforcement powers. The government now has two months to respond to the Committee’s findings.

Which human rights are at risk from AI?

The report identifies the following among the principal rights at risk from AI:

• Protection from discrimination. Bias can become embedded in AI systems during development, as models are trained on large datasets that may reflect inherent human biases or be unrepresentative. For example, AI tools have been used for profiling in prisons to assess individuals’ security categorisations and risks of violence, which has been found to disproportionately affect people of colour. In the workplace, AI systems have used emotion recognition to flag workers for disciplinary action without reasonable cause.
• Right to a fair trial. Where public bodies are not transparent about their use of AI, individuals may be unable to understand, or effectively challenge, AI-related decisions concerning them. For example, police forces have used live facial recognition technology to scan and identify people in public places without their consent, and transparency is crucial to properly challenge such identification evidence in court.
• Protection of property. Data used to train AI is often gathered from online sources and may include personal data obtained without appropriate consent. The use of copyright-protected material, such as artworks, to train AI also raises intellectual property concerns.

The JCHR’s recommendations

The report describes the UK’s existing legal framework for AI as fragmented and difficult to navigate, with the majority of legislation designed for specific harms and sectors, leaving gaps in human rights protection. Most legislation applies only at the point of deployment of an AI system, meaning responsibility lies ‘downstream’ with the businesses that use AI, rather than with developers, who are often excluded from liability in supply contracts.

The JCHR’s main legislative recommendations are:

• A risk-based regime. AI systems posing the greatest risk to human rights would face stricter legal requirements than those classed as lower risk. This approach would be similar to that employed by the EU AI Act (Regulation (EU) 2024/1689).
  Outright bans on AI uses which are incompatible with human rights. This could include the use of AI for profiling techniques, and the use of biometric data.
• Liability across the supply chain. Obligations would apply to all designers, developers and deployers of AI, according to their role and the level of risk posed. It was recognised that large developers will often be better placed and resourced than deployers to identify risks and prevent AI harms.
• Mandatory transparency when using AI systems with the potential to significantly impact individuals and groups. In these cases, organisations would be required to specify when and how AI is being used, making it easier for affected people to challenge decisions.
• Stronger safeguards on automated decision-making. The Committee stressed that simply having a ‘human in the loop’ does not amount to meaningful human involvement in an automated process.
• A single, independent, statutory AI oversight body.

The report signals a possible move towards a structure more closely aligned with the EU, with obligations that would also reach AI developers, not just deployers and users. As the UK currently has no equivalent to the EU AI Act, some form of alignment can reasonably be expected. The JCHR also emphasised that regulation should be proportionate and should not place unjustified burdens on businesses, especially those using the lowest risk AI systems, and should not overly restrict innovation and growth, recognising that AI also has numerous benefits.

What next?

The government must respond to the JCHR’s findings by mid-November. Given that the government’s approach to AI so far has focused on the growth benefits and a preference for a ‘regulator-led’ approach to governance, the report could signal a significant shift. In a related move, the Law Commission of England and Wales has launched a consultation, open until January 2027, on whether the current product liability regime is fit for purpose in relation to digital products and AI. A key proposal is amending the statutory definition of ‘product’ to include software and AI systems, which would bring the UK into line with the new EU Product Liability Directive (Directive (EU) 2024/2853), which applies from 9 December 2026.

The message is clear: although the UK does not yet have an AI-specific law, pressure to introduce one is growing, particularly to align the UK with related provisions at the EU-level. Businesses who build governance frameworks into their use of AI now are likely to be best prepared for what follows.