Tech4Humanity AtlasGround ZeroCurrent ThemesFuture ResearchGalleryLive Q&ASearch
Ground Zero programme

Extreme AI Effects

1 authored topic chapters · 6 studies with linked canonical stories.

Topics

Performance Extremes

Performance Extremes brings together Extreme Gains, Negative Responders, The Impact of Drugs, High Stakes Use Cases, and Dependency & Atrophy, and related investigations. Within the Extreme AI Effects Theme, these studies examine a shared problem from different human, technical and institutional perspectives. The Topic makes their relationships explicit so that methods, risks, discoveries and unre

  • SUB-0007 — Extreme Gains: Which user, task and AI combinations produce extreme positive performance outliers?
  • SUB-0008 — Negative Responders: Which signals predict that an individual will experience net harm from a given AI assistance regime?
  • SUB-0009 — The Impact of Drugs: How do common stimulants, depressants and prescription agents alter optimal AI assistance thresholds?
  • SUB-0010 — High Stakes Use Cases: Which extreme-effect patterns most threaten safety-critical decision quality?
  • SUB-0011 — Dependency & Atrophy: At what cumulative exposure does AI support begin to erode independent capability?

Studies & stories

GZ-STORY-007

Extreme Gains

The dashboard was green, but Daniel Okafor did not trust it. The team had finished faster with AI, yet nobody could say whether they had learned more, understood less, or simply moved the verification

Read the story →
GZ-STORY-008

Negative Responders

At 8:10 on Monday morning, Grace Liu watched two people receive the same AI advice and move in opposite directions. One became more capable. The other became quieter, less certain and increasingly dep

Read the story →
GZ-STORY-009

The Impact of Drugs

The meeting began with a number everyone liked and a question nobody could answer. Output had risen. Complaints had fallen. But Tom Bennett asked what had happened to judgement, agency and the ability

Read the story →
GZ-STORY-010

High Stakes Use Cases

The dashboard was green, but Aisha Rahman did not trust it. The team had finished faster with AI, yet nobody could say whether they had learned more, understood less, or simply moved the verification

Read the story →
GZ-STORY-011

Dependency & Atrophy

At 8:10 on Monday morning, Ben Carter watched two people receive the same AI advice and move in opposite directions. One became more capable. The other became quieter, less certain and increasingly de

Read the story →
GZ-STORY-021

Extreme Performance Zones

The meeting began with a number everyone liked and a question nobody could answer. Output had risen. Complaints had fallen. But Inez García asked what had happened to judgement, agency and the ability

Read the story →