UK draft guide asks government staff to stop thanking AI chatbots

Employees remain accountable for AI-assisted content and decisions, rather than handing responsibility to the tools they use.

The UK has issued an unfinished guide for government staff on responsible AI use, emphasizing safe, fair and sustainable practice. It directs employees to observe rules set by their departments and select a model suited to the task; examples include Gemini Flash instead of Gemini Pro and GPT Instant instead of GPT Thinking. The draft says staff should check AI output for accuracy and equitable treatment, and stresses that employees remain responsible for AI-assisted content and decisions. It also presents skipping thank-you messages to chatbots as part of efforts to reduce AI's environmental impact.