TO THE EDITOR:
Ricardo Martinez's comments are excellent ("What is the acceptable death rate in autonomous vehicle accidents?" April 1), but he misses the point of the capabilities of artificial intelligence.
Today's AI systems are incapable of guaranteeing no road deaths from autonomous vehicles. Another 10 years or more of research are needed until AI will approach near human intelligence.
In the interim, if every vehicle on the road were autonomous, we might reduce annual U.S. traffic deaths from 40,000 to, say, 1,000.
From a physician's perspective, are those 39,000 lives worth saving at the expense of 1,000 deaths caused by AI errors? Can government find a way to stop litigation against the 1,000 deaths for the benefit of the 39,000 lives saved? Or is American culture too challenging for this calculation to be viable?
I would vote to save the 39,000.
ALAN MARTIN, Founder and director, IQZero
Bradenton, Fla. IQZero is a think tank focused on AI, future mobility and solutions beyond digital.