Future of Work · 11 min · Suggested 16+; not age-rated
Building AI that Doesn’t Go Rogue
Deni Darenberg
Deni Darenberg explores building AI That Doesn't Go Rogue. The conversation highlights AI safety and alignment, with ideas about AI products. Viewers can use the learning outcomes and questions below to reflect on the talk and consider what to try next.
Key moments
- 01
AI needs reliable information, not just large amounts of it
- 02
Build in limits and a way to involve a human
- 03
Teams deploying AI are responsible for ethics and context
For teachers
Suggested curriculum links
Possible Gatsby Benchmark links
- BM4
Suggested alignment only. Watching this talk does not on its own meet a benchmark; use it within a planned careers activity. BM4 requires a planned subject-to-career link, not video viewing alone. BM2 needs current credible labour-market information; BM7 needs actual exploration of pathways. These are potential supporting links, not evidence of benchmark completion.
Ask OffScript+
Ask the speakers anything
Not sure what to ask? Start here.
Explore what our speakers have learnt along the way.
Find your way
Get ahead
Real life, real lessons
Answers come only from checked talk transcripts and link to the moment in the video.