CableLabs futurist sees limits to artificial intelligence
ATLANTA – As a futurist for the cable industry, Bernardo Huberman sees huge potential for artificial intelligence and machine learning in many realms, but that doesn't mean he would trust smart machines to take on all the tasks that human beings handle right now.
Take flying, for example. Despite his enthusiasm for AI, Huberman, a fellow at CableLabs and VP of the organization's Core Innovation Team in Silicon Valley, would not allow an AI-equipped machine to fly passenger plane all by itself, at least not one carrying him. He'd much rather have a real, flesh-and-blood pilot. (Ed note: Us, too!)
"You don't want to fly in a plane with no one in the cockpit," Huberman said in a fireside chat here at the SCTE Cable-Tec Expo show last month. "I like to know there is someone competent in the cockpit when I fly."
Nor would Huberman let a smart machine drive a bus or train anywhere, except possibly under very limited, controlled circumstances where little harm could be done. Likewise, he's not quite ready to trust fully autonomous automobiles. While he endorses the idea of self-driving vehicles and looks forward to seeing such connected cars tooling down the road in the future, he believes many things must still be worked out – including the technology enabling autonomous autos, the roadway systems supporting them and the regulations covering them – before the cars can actually hit the streets in any great numbers.
"The rollout will be a lot slower than predicted," he predicted. "It will require tremendous caution and regulations."
Finally, Huberman is certainly not ready to let AI-run national defence systems launch nuclear attacks on other countries. He discussed the famous case of the late Stanislav Petrov, a lieutenant colonel in the old Soviet Union's air defence corps who single-handedly prevented an all-out nuclear war from breaking out with the U.S. in September 1983 by overruling his nation's primitive AI system.
Specifically, with Soviet-American tensions running high in the wake of the Russians' downing of a Korean passenger plane just three weeks earlier, the Soviet Union's satellite early-warning system reported that the U.S. had just launched a nuclear missile towards Russia and had up to five missiles more on the way. However, rightly judging that report to be a false alarm, Petrov decided to go against Russian military protocol and called off a retaliatory nuclear strike against the Americans and their NATO allies.
"He thought it very unlikely it [a U.S, nuclear attack] would happen like this and the U.S. would launch just five nuclear missiles," Huberman said. "That's something a machine couldn't figure out… but he admitted there was a 50% chance that he was wrong."
Even with such caveats, though, Huberman is quite optimistic about the future of both AI and machine learning in numerous fields and applications, especially ones like distance education, however, he added, "I totally believe we cannot replace humans in the education process.”
