ChatGPT (‘Conversational Generative Pre-Training Transformer’) and the release of GPT-4 is dominating the news. Some firms are investing in it, and some are banning their staff from using it. Risks include breaching confidentiality, advice which is entirely plausible but wrong, copyright breaches, and breaches of data protection legislation. It has been demonstrated that it can pass the US bar exam and the Gazette reported that the previous version GPT-3 had a reasonable attempt at the Solicitors Qualifying Examination.
Links to guidance from The Council of Bars and Law Societies of Europe, newly updated Guidance on AI and data protection from the Information Commissioner’s Office (15 March 2023), and a National Cyber Security Centre blog, ChatGPT and large language models: what’s the risk? are on www.legalrisk.co.uk/News.
Firms which are considering using third party vendors of AI services will need to consider carefully the implications in relation to information security.
‹ Back to Publications