Just published something about how we’re designing GOV.UK Chat, a product that’s currently in a private beta across some GOV.UK pages
https://insidegovuk.blog.gov.uk/2024/11/28/how-were-designing-gov-uk-chat/
https://insidegovuk.blog.gov.uk/2024/11/28/how-were-designing-gov-uk-chat/
Comments
(* this bit is really hard!)
Interesting that nearly 1 in 5 people using GOV.UK Chat do not realise that it can give inaccurate answers.
Any sense of whether this 20% is likely to be biased towards those most vulnerable?
https://tomloosemorework.wordpress.com/2024/04/18/ai-and-the-cost-of-failure/
And - how does the gen AI content stand up against GOV.UK's styleguide.