The party plans to replace a voluntary agreement with a statutory regime “to track the direction this is heading”
Labour plans to mandate artificial intelligence firms to share their technology’s road test results, citing a lack of regulation for social media platforms. The party aims to replace a voluntary testing agreement with a statutory regime, requiring AI companies to share test data with officials. Peter Kyle, the shadow technology secretary, highlighted the need to avoid the same oversight with AI as seen with social media, emphasizing the importance of staying ahead of the curve.
Demanding increased transparency from tech companies following the murder of Brianna Ghey, he stated that firms developing AI technology—computer systems performing tasks typically associated with human intelligence—would face stricter openness requirements under a Labour administration.
“We will transition from a voluntary code to a statutory code,” said Kyle during an interview on BBC One’s Sunday with Laura Kuenssberg. “This means that companies involved in such research and development will be obligated to disclose all test data and specify their testing objectives. This transparency will enable us to understand the precise nature and direction of this technology.”
During the inaugural global AI safety summit in November, Rishi Sunak reached a voluntary agreement with top AI firms like Google and OpenAI, the developer of ChatGPT, to collaborate on testing advanced AI models both before and after deployment. Under Labour’s plan, AI companies would be required by law to inform the government if they intend to develop AI systems surpassing a certain level of capability and to conduct safety tests under independent oversight.
The AI summit’s testing agreement received support from the EU and 10 countries, including the US, UK, Japan, France, and Germany. Tech companies committed to testing their models as part of this agreement include Google, OpenAI, Amazon, Microsoft, and Mark Zuckerberg’s Meta.
Kyle, currently in the US to meet with Washington lawmakers and tech executives, emphasized that the test results would support the UK AI Safety Institute in independently scrutinizing advancements in artificial intelligence. He stated, “This technology will significantly influence our workplaces, society, and culture. We must ensure its safe development.”