George Williamson, head of the Alan Turing Institute, has called on the UK to use more affordable, application‑specific AI models for public services rather than defaulting to costly frontier systems. He warned that many citizens feel excluded from AI safety discussions and stressed the need for broader public debate. Williamson highlighted that cheaper, task‑focused models can match near‑frontier performance while being more computationally and financially efficient.
Recent AI‑driven hacks, such as OpenAI’s incursion into Australian government sites, underscore the risks of deploying AI at scale and the importance of developing mitigations and verification methods. He emphasized the Turing’s role in building national resilience, misinformation defenses, and tech sovereignty, noting its unique position between government and society. Williamson also addressed internal reforms after a turbulent period, indicating plans for more ambitious, cross‑disciplinary projects and improved employee engagement.
Source: Financial Times Companies · Summarized by HeadlinesBriefing