Hong Kong's metro puts others to shame: It's one of the most profitable train networks in the world. It's on time 99.9 per cent of the time. It's always improving — and it's controlled by some very clever artificial intelligence.
An article in the latest issue of New Scientist introduces us to the algorithm that's responsible for the incredible task of repairing and maintaining the system. That means assigning 10,000 employees to take care of more than 2500 engineering jobs every week — an insanely complicated puzzle that would normally take a panel of humans two days of strategising and planning. Instead, their bot can calculate the most efficient assignments and adapt to new information in seconds.
In the end, smarter assignments mean that the employees have more time to finish their nightly jobs, and according to New Scientist's Hal Hodson, the time bumper saves the system $800,000 a year. There's just one problem: People don't like being kept in the dark about the grand plan. It's still difficult for us humans to trust a machine to do our jobs — no matter how advanced. [New Scientist]
Picture: AP Photo/Vincent Yu