2:36Research has to happen during deployment
Rachel Coldicutt explains why AI evaluation must continue during use, asking who is excluded when a tool fails and whether it works in each new context.
Roles are listed as credited at the 2026 summit. Passage assignments are based on transcript context and remain unconfirmed.
2:36Rachel Coldicutt explains why AI evaluation must continue during use, asking who is excluded when a tool fails and whether it works in each new context.
3:05Rachel Coldicutt proposes shared maintenance and a culture of safety that continues beyond the pilot, the grant and the launch checklist.
1:32Drawing on research in southern India, Rachel Coldicutt describes small enterprises building for specific languages, needs and levels of access.
Broad categorisation; not exhaustive technical assessment.
Research account without methods or detailed results here.
Hypothetical time horizons; no universal decision rule.
Analytic account; magnitude not measured.
Illustrative percentages are not study estimates.
Suggested quarterly cadence is not a tested universal standard.