Have AI Got News for You
summary
Claims that LLMs can produce FT‑quality first drafts but not final copy; an internal AI playground is built and rolling out to Slack for journalists to test with FT content; retrieval uses retrieval augmented generation on ~15 selectable 300‑word segments; prompting techniques—few‑shot examples (3–5) role priming and Chain of Thought—improve factuality and structure but hallucinations and lack of emotional thrust persist, so the tool speeds early drafting yet fails to replace editorial judgment.
AI-written summary · about
chapters (9)
- 0:00AI in journalism
- 1:26Accelerate AI team
- 2:54Large language models
- 5:26Prompt engineering
- 7:53AI Playground tool
- 9:20Live demo: AI Playground
- 12:11Prompting techniques
- 18:18LLM limits in journalism
- 19:46Future of AI in news
related talks
59mIt AI n't What You Think!
27mStereo-tAIping: Who Gets to Write the Story?
25mAI vs. AI: How to Test AI Systems Using AI
09mRethinking AI: Building What Matters, Not Just What's Possible
40mBeyond the Hype: Realistic Expectations of AI
55mAI-First Testing
01h 39mMastering LLMs: From Prompt Engineering to Agentic AI
18mAI Quality Offensive — Content Quality without Slop and without Prompts