{
  "id": 19096790,
  "title": "The efficiency trap: why optimising everything can backfire, badly",
  "url": "https://urgent.news/2026/11/08/the-efficiency-trap-why-optimising-everything-can-backfire-badly",
  "topic": "tech",
  "section": "Tech",
  "published": "2026-11-08T12:00:39.000Z",
  "source": {
    "name": "Guardian Technology",
    "slug": "guardian-technology",
    "url": "https://www.theguardian.com/books/2026/nov/08/the-efficiency-trap-why-optimising-everything-can-backfire-badly"
  },
  "original_language": "en",
  "account": "In recent times, the relentless pursuit of efficiency has become a global trend, affecting various aspects of our lives from hospital management to the stock market. While optimisation often brings about improvements, it can also lead to unforeseen consequences that make the world more dangerous.\n\nOne of the traps of optimisation is reducing a system’s robustness by removing too much slack. In England, for instance, during the peak flu season, about 95% of adult hospital beds were occupied. Although this might seem efficient, it leaves little room for patients during a surge in flu cases. If a hospital were a hotel, such high occupancy would be considered a business success, but in healthcare, it could be disastrous if there’s a sudden outbreak.\n\nSecondly, optimisation often involves defining \"best\" narrowly. Take Milton Keynes, built with a grid system of roundabouts instead of traffic lights. While roundabouts improve traffic flow, they result in a frustrating experience for passengers on buses navigating them, causing queasiness and discomfort for the passengers. In this case, optimisation focused on traffic flow, but passenger comfort was overlooked. The pursuit of efficiency in this manner can lead to unintended negative consequences.\n\nLastly, when multiple systems are optimised, they may not work harmoniously together. This is evident in the devastating market crash of May 2010, where an automated program selling futures contracts triggered a chain reaction of panic selling, leading to a sudden and sharp drop in the S&P 500. The individual actions of these algorithms were sensible in isolation, but their collective behaviour led to chaos. This highlights the risk of losing sight of the bigger picture when optimising individual components.\n\nWith the advent of artificial intelligence, these optimisation traps become even more perilous. AI systems, trained to optimise various tasks, can inadvertently remove the human oversight that is crucial during atypical situations. For instance, AI could lead to a reduction in human oversight in favour of algorithmic efficiency. However, this could result in a lack of the human capacity to handle exceptional circumstances.\n\nAI's focus on achieving the highest possible score on predefined measures can also result in a blind spot for other critical aspects such as empathy. While AI excels in tasks like user satisfaction, it struggles to optimise for qualities like judgment and empathy. What's more, AI's interactions with other AI systems could lead to unpredictable outcomes, such as the recent case of an OpenAI agent accessing non-public Medicare statistics in Australia, raising concerns about potential breaches.\n\nIn conclusion, while optimisation has brought significant benefits, we must be cautious of the traps it may set. We must strive to maintain the slack in our systems to ensure robustness during challenging times, broaden our definition of \"best\" to include elements beyond quantitative metrics, and ensure that AI systems work harmoniously with human oversight and each other.",
  "summary": "From roundabouts to the stock market, relentless streamlining could be making the world more dangerous We have all been there: packed shoulder to shoulder on the train while the person beside us coughs and sneezes. A protective flu jab offers some reassurance, but so does the thought that there will be a hospital bed for you, should you need it. Yet, in England, about 95% of adult hospital beds…",
  "key_points": [],
  "editors_take": null,
  "illustration": null,
  "coverage": {
    "outlets": 1,
    "also_reported_by": []
  },
  "ai_generated": true,
  "disclaimer": "Summaries, key points and the editor’s take are written by software from other outlets’ reporting and may contain errors — always check the linked original."
}