��ҹ����1000����

What is Curiosity Engine?

Curiosity Engine is an AI chatbot developed by New Scientist and technology company Miso. Its goal is to have interesting science-filled conversations rooted in New Scientist’s rich archive of stories.

How does Curiosity Engine work?

Curiosity Engine uses a type of artificial intelligence called a large language model. These AIs are trained using a huge amount of data, allowing them to respond to many kinds of queries. However, they are also prone to hallucinations – a term used to describe when AIs confidently assert incorrect statements. To limit this, Curiosity Engine uses a technique called Retrieval-Augmented Generation (RAG), which aims to force the LLM to stick to the facts in the New Scientist archive. This works often, but like with all AI, sometimes it doesn’t, so we recommend regularly checking the source articles for more information.

How up to date is Curiosity Engine?

Curiosity Engine has access to the latest relevant articles published by New Scientist within 10-15 minutes.

What should I do if I spot Curiosity Engine making a mistake?

You can give quick feedback about Curiosity Engine using the thumbs up and thumbs down buttons at the bottom of an answer. If you would like to give further feedback, please send us an email at curiosityengine@newscientist.com.

Who can see my conversations?

Anonymised questions and answers are stored in a log that select members of the team can access. By “anonymised” we mean that we do not collect any information about who is asking the question beyond their subscription status. So, for example, a New Scientist editor can review the answers logs and see that a subscriber asked a question, what the answer was, any feedback the user provided, whether they asked a follow up question and so on.