發生了什麼事
The Law Society Gazette reports that the Joint Committee on Human Rights has urged the UK government to introduce an AI-specific legal framework covering all sectors. The committee says existing laws are fragmented and place too much responsibility on deployers rather than developers, and it recommends a precautionary approach and a single AI regulator.
The Law Society Gazette reports that the Joint Committee on Human Rights called for an “AI-specific legal framework” applying across society and said regulation should take a precautionary approach because of potential risks to human rights.
According to the report, the committee’s Human Rights and the Regulation of AI report found that UK laws covering AI are fragmented and generally focus liability on those deploying systems rather than on developers. The committee argues that this may leave risks with the actor least able to prevent them.
The committee also cited evidence that AI use in court proceedings could threaten the right to a fair trial under Article 6 of the European Convention on Human Rights. It recommends that the government introduce a new AI bill, promised in the 2024 King’s Speech, broadly aligned with the Council of Europe’s non-binding framework convention.
Committee chair Alex Sobel MP said the UK and other countries are not currently prepared to deal with AI’s consequences and called for a single regulator able to set policy, monitor performance and enforce compliance.
為什麼這很重要
The committee’s position would move UK toward lifecycle-wide accountability, potentially assigning clearer duties to developers as well as deployers. That could affect systems used in courts, public services and other high-impact settings. The Law Society Gazette reports that the committee linked AI-generated risks to human rights, including the right to a fair trial. The report does not establish that these harms have occurred, and it does not show that the government has accepted the recommendations.
A framework covering the full AI supply chain could change how responsibility is allocated between model developers, system integrators, deployers and public authorities. The source’s central concern is that current arrangements may not target the party best placed to identify and mitigate risks.
The implications are particularly significant for high-stakes uses such as courts, where errors or fabricated outputs could affect procedural fairness. However, the supplied report contains the committee’s assessment and recommendations, not independent testing of specific AI systems or evidence that a particular case was compromised.
The recommendation also signals a potential shift from voluntary or sector-by-sector governance toward a more coordinated statutory regime. Its practical effect remains unknown until the government publishes concrete legislation.
互動機制：它實際上是如何運作的
以互動方式探索這項發展背後的基礎技術。
Why can ethical evaluation not be reduced to one model score?
接下來看什麼
The key next step is whether the UK government introduces the AI bill referenced by the committee and what duties, enforcement powers and institutional responsibilities it contains. Important unknowns include the bill’s timetable, whether it would create a single regulator, how it would interact with existing sector-specific rules, and whether developers would face direct liability. The supplied report gives no access or pricing details because this is a policy proposal, not a product announcement.
Watch for a government response and any published AI bill, including its scope, definitions, duties for developers, liability provisions and enforcement mechanisms.
The proposed single regulator would require clarification about its relationship with existing UK regulators and whether it would have binding powers, technical capacity and sufficient resources.
The source does not provide a timetable, draft legislation, implementation plan or evidence that the government has endorsed the committee’s recommendations.