Audrey Pariente · PhD researcher, Sheffield Hallam University
The Risks of AI to Strategy
I study how generative AI quietly narrows strategic decisions long before they are made, and how leaders and organisations can keep human judgement in charge.
01The big question
How does generative AI reshape strategic decision-making by altering judgement, cognitive authority and responsibility, and how can these changes be responsibly governed?
Generative AI gives leaders access to more information than ever. But its outputs favour the dominant patterns in its data and tend to erase rare information: weak signals, contrarian views, unusual events. Strategy, however, lives in exactly that rare information. The result is a paradox: more insight, less foresight.
02What I work on
An article-based PhD that follows generative AI from the individual leader, to the organisation, to the wider institutional field.
25 risks to strategic decision-making
How GenAI narrows what leaders search, see and weigh.
GovernanceDisclosed versus practised governance
What firms say about AI, and what they actually control.
JudgementPhronetic leadership
Own the judgement, pause before commitment, stand behind the choice.
ContextEducation and the Global South
Where limited AI capability shapes how the technology can be used.
03Recent work
- 2026-05Best Paper Award, DIGEST 2026“Phronetic Leadership” won Best Paper in Track 3, Inclusion, Equity and Digital Justice, at Sheffield Business School.
- 2026More Insight, Less ForesightAccepted in Electronic Markets (open access), with D. Wagner and A. Richter.
- 2026Leading Under AI DeprivationAccepted chapter in a De Gruyter handbook on generative AI and the Global South.
- 2026-04Confirmation of doctorateResearch programme confirmed at Sheffield Hallam University.
04Background
Twenty years of practice before the PhD, and still in practice alongside it.
That path, from teaching languages to leading communications for a semiconductor leader, is why my research asks practical questions about how decisions are really made.