Petra Stock
The European release of yet another artificial intelligence (AI) chatbot has run aground over privacy concerns.
Google’s AI chatbot Bard cannot be released as planned in the European Union until the tech company can demonstrate its compliance with privacy legislation, according to reporting in Politico.
Bard is a large language model similar to ChatGPT, a form of generative AI trained on text data largely drawn from the internet as well as user prompts.
Deputy Commissioner Graham Doyle from the Irish Data Protection Commission says the privacy regulator “had not had any detailed briefing nor sight of a data protection impact assessment or any supporting documentation at this point,” according to the article.
In March, Italy’s privacy regulator temporarily banned OpenAI’s ChatGPT over concerns regarding compliance with privacy laws such as the General Data Protection Regulation (GDPR).
In Australia, similar concerns have been flagged by Rachel Dixon, Deputy Commissioner of privacy and data protection at the Office of the Victorian Information Commissioner.
Dixon told Cosmos it’s not appropriate for governments to be using large language models in their work due to the risks of conflict with government responsibilities to protect personal information under Australian laws.
The Australian Government’s eSafety Commissioner warns Google Bard is “experimental and has certain limitations” noting the model may display inaccurate information, offensive statements and may infringe intellectual copyright by incorporating content without permission.