Anthropic to use Claude chats to train future AI models

Anthropic is updating its privacy policy to incorporate the use of user conversations with its Claude AI chatbot as training data for future large language models (LLMs), unless users opt out.

The change, which takes effect on 8 October, represents a shift from the company's previous policy of not training its generative AI models on user chats. Anthropic stated that incorporating data from real-world interactions will provide valuable insights into which responses are most useful and accurate, enabling them to develop a better version of Claude over time. 

Meanwhile, The Wall Street Journal (£) reports that Meta plans to use conversations people have with its AI chatbot to personalise ads and content on its Instagram and Facebook platforms, effective from 16 December. The new policy marks a fresh approach to digital privacy and helps fund the company’s expensive AI efforts. While users cannot opt out of this processing, the company confirmed that chats involving sensitive topics, such as religious views, health, or political opinions, will be automatically excluded. The policy will not initially apply to users in the UK, the EU, or South Korea.

£ - This article requires a subscription.

Training Announcement: Freevacy offers a range of independently recognised professional AI governance qualifications and AI Literacy short courses that enable specialist teams to implement robust oversight, benchmark AI governance maturity, and establish a responsible-by-design approach across the entire AI lifecycle. Find out more.