Optimising systems for efficiency can make them fragile, a Guardian columnist argues
Overview
A Guardian columnist argues that optimising a system for efficiency can weaken its robustness, narrow what counts as "best", and create conflicts between automated processes.
The column cites England's hospital beds, which were about 95% occupied at last year's flu peak, and the 2010 US flash crash, in which an automated program's sale of 75,000 futures contracts triggered market-wide turmoil. It warns that AI, which is built by optimisation, could make these traps more dangerous, including through interactions among swarms of AI agents.
Written by AI from the articles below · updated Nov 8, 7:13 AM ET
Check the sources:
Article timeline
The articles in this story. Times are ET.
- The Guardian · AINewsThe efficiency trap: why optimising everything can backfire, badly
AIColumnist argues that optimising a system for efficiency can weaken its robustness, narrow its definition of "best", and cause conflicts between automated processes. The article cites England's hospital beds, which were about 95% occupied at last year's flu peak, and the 2010 US flash crash in which an automated program's sale of 75,000 futures contracts triggered market-wide turmoil. It warns that AI, which is built by optimisation, could make these traps more dangerous, including through interactions among swarms of AI agents.
Heat Trend
Not enough continuous observations to show a trend yet.