The UK government has released draft guidance titled Using AI ethically and sustainably, aimed at regulating how staff interact with artificial intelligence tools. The document advises employees to avoid thanking chatbots, noting that such interactions are unnecessary and contribute to the environmental impact of AI usage.
To improve efficiency, the guidance instructs staff to use short, clear, and direct prompts. It also recommends selecting lightweight models for specific tasks, citing Gemini Flash as an alternative to Gemini Pro and GPT Instant as a substitute for GPT Thinking. Employees are reminded to use the most appropriate tool for each job.
The document emphasizes that users must verify all AI-generated output for accuracy and potential bias, ensuring that groups are treated fairly and without discriminatory language. Staff are held personally responsible for any content or decisions made using AI, as they cannot shift accountability to the software if errors occur.
Additionally, the guidance warns that AI usage policies may vary across individual government departments. It suggests that employees should consult their specific departmental guidelines regarding approved tools and approaches, reflecting the diverse nature of tasks performed within the government.
