US board drafts AI ethics guidelines for tech cos

■ A governable AI, that can stop itself if it detects that it’s about to cause unnecessary harm, is desired.

people being treated unfairly, the board said, even if they’re not necessarily in life-and-death situations.

It also wants to check with the military, that its data sources were neutral.

If not for all situations, bias could be useful for targeting key combatants or minimising civilian casualties, the draft mentioned.

The draft calls for a ‘governable AI’ that can stop itself if it detects that it’s about to cause unnecessary harm and control itself in time. It also suggests switching to a human operator if need be.

Highlighting the risks involved in having fully automated weapons, the board aims at reducing the chances of AI becoming uncontrollable.

The draft includes a call for ‘traceable AI’ output that lets people see how an AI system reached its conclusion. While the draft is promising, the challenge is implementing it.

It is easy to promise more accountable and trustworthy AI, but it is another thing to ensure that every military branch follows those ideals with every project.

As Defense One observed, though, the department may have an advantage over tech companies in that it’s starting with a relatively blank slate. It doesn’t have to make exceptions for current AI projects or else rethink its existing strategy — the guidelines should be there from day one.