Should we trust AI?
As the technology becomes more foregrounded in all aspects of life, Shaun Prescott asks: where do we draw the line?
Artificial intelligence, once only of pressing import to science-fiction fans and videogame enthusiasts, is increasingly dominating tech news headlines. It’s been tipped as the answer to many modern conundrums, whether it be streamlined consumer services online or as a means to combat the dissemination of ‘fake news’. But during a time when tech companies like Microsoft are investing heavily in AI, it’s worth wondering whether it’s a good idea to hand over so much power to systems. Indeed, is our world even prepared for the possibility that entire workforces could be wiped out by even more AI-led automation?
Even Microsoft founder Bill Gates is sceptical about AI, while Elon Musk seems vaguely scared of it. “I think we should be very careful about artificial intelligence,” Musk said at a symposium in 2015. “If I were to guess at what our biggest existential threat is, it’s probably that. So we need to be very careful with the artificial intelligence. Increasingly scientists think there should be some regulatory oversight maybe at the national and international level, just to make sure that we don’t do something very foolish. With artificial intelligence we are summoning the demon.”
“Summoning the demon” is a far cry from “getting a computer to change your internet plan”. Whether that’s hyperbole or not, Gates, Musk and even Stephen Hawking are not alone among authority figures when it comes to fears regarding AI. According to Google researcher Chris Olah, speaking about a paper he contributed to on the subject, “It’s essential to ground concerns in real machine-learning research, and to start developing practical approaches for engineering AI systems that operate safely and reliably.”
These concerns range from preventing cleaner bots from causing electrical fires through to deeper concerns. As far as the latter goes, AI’s effect on employment is of growing interest, especially on low-income and low-skill employment. The White House released a report late last year entitled ‘Preparing for the Future of Artificial Intelligence’, which did not shirk from the question of whether jobs would be lost — indeed, it’s very likely they will be, en masse. But it’s not just low-skilled work that’s on the chopping block: according to a report by Fortune, specialists have pointed to the likelihood that commercial airlines may one day be piloted by AI. That doesn’t sound scary at all.
Still, it’s hard to deny that AI’s application to problems which are difficult to address by mere humans is worth advancing. Fake news, for instance, has proliferated to the extent that gatekeepers like Facebook cannot possibly police it with human moderation: can deep learning and algorithm-based vetting keep it from being distributed too widely and too dangerously? That indeed, is what many are positing as the answer. Better that than a computer operating heavy vehicles or driving your children to school.