A cross-party group of UK Members of Parliament and peers has called for new legislation to address the human rights risks posed by artificial intelligence, arguing that current laws are insufficient. The Joint Committee on Human Rights (JCHR) published a report stating that no country, including the UK, has a legislative and regulatory approach to AI that is fit for purpose.
The committee cited concerns that AI can perpetuate biases if trained on datasets containing undesirable material, and that it can be used to create and spread misinformation. Specific examples of AI-driven human rights abuses mentioned in the report include the creation of sexualised images of women and girls and the scanning of people's faces without consent.
The JCHR advocates for the establishment of a "single, independent AI oversight body... on a statutory basis," noting that the current legal framework is "fragmented and difficult to navigate." The proposed AI Bill should classify different risk levels for AI systems and impose more demanding obligations on higher-risk models, with responsibilities placed on all stages of the AI lifecycle. The report also suggests that certain AI uses, such as subliminal techniques or inappropriate profiling, should be prohibited outright.
Alex Sobel MP, the committee's chair, described AI as a technology with the potential to transform lives for better or worse, but warned that society is currently unprepared for its consequences. The report comes amid broader public anxiety about AI's challenges, with some AI researchers expressing fear for humanity's future. Anthropic, a rival AI firm, stated it is building models with strong safeguards.