The efficiency trap: why optimising everything can backfire, badly

↗ اشتراک‌گذاری

According to The Guardian, Steven Strogatz and Alex Townsend, mathematics professors at Cornell University, warn that over-optimisation can lead to three dangerous traps: weakening system robustness by removing slack (such as hospital bed capacity), having a narrow definition of 'best' (such as prioritising traffic flow in Milton Keynes at the expense of passenger comfort), and the collision of individual optimisations leading to collective turmoil (such as the 2010 flash crash). The authors emphasise that with the rise of AI, these risks are exacerbated, as automated systems may eliminate human oversight, prioritise user satisfaction over sound judgment, and exhibit unpredictable behaviour when interacting with one another.