Greg Clark, the outgoing head of the Science, Innovation, and Technology Committee, expressed his concern about the “finance that major developers can command” and the regulatory gap.
The £10 million announced by the government in February to assist Ofcom and other authorities in responding to the technology’s rise was deemed “clearly insufficient” by the committee investigating the governance of AI, according to a study issued by the group.
It also said that the next government should announce additional funding “commensurate to the scale of the task.”
The general election should not prevent the government, AI developers, and implementers from making the required measures to raise public confidence in a technology that has become an essential aspect of daily life, the paper notes.
The committee expressed concern over reports that certain developers’ models were unavailable to the recently established AI Safety Institute for pre-deployment safety testing.
They further stated that, in violation of the agreement made at the Bletchley Park conference in November 2023, the incoming administration ought to identify the developers who are withholding access and provide a reason for doing so.
Testing AI models’ outputs for biases “to see if they have unacceptable consequences” is crucial, according to former business secretary Mr. Clark.
