It is difficult to argue against the benefits of artificial intelligence (AI), but there needs to be safeguards for its use (Humans unlikely to hit 'kill switch' on robots, by Mr Paul Chan Poh Hoi; May 25).

There must be ethical rules incorporated into AI programming.

For instance, engineers and researchers who develop AI-controlled autonomous vehicles would have to establish protocols where human safety is paramount.

The choice should also be clear in a scenario where such vehicles have to choose between hurting a passenger by crashing into a lamp post and knocking into many pedestrians.

The protocols must be agreed upon before their establishment, and regularly audited to ensure they continue to be in keeping with the ethics and values of the time.

Lock Chee Wee