Much attention has been paid to the existential risks of AI, many of which relate to so-called AGI (artificial general intelligence) technologies, which are yet to be developed. Eighteen months into this ‘AI Summer’, these existential risks have yet to materialise, and experts disagree on when — and indeed, if — we will see AGI. Existing and imminent generative AI systems and large language models (LLMs) like ChatGPT, Llama, Palm and Claude present immediate opportunities and challenges. They require a sound AI infrastructure.
Our research shows flaws that could cause model and business collapse, due to lack of data transparency and poor data-sharing practices threatening rights and livelihoods. If the UK is to benefit from the extraordinary opportunities presented by innovations in this field, it must look beyond the hype cycle and return to the data fundamentals that are so essential for building a robust data and AI ecosystem, and a strong, equitable market that benefits everyone, and minimises harm.
