Sunday, April 01, 2018 by Vicki Batts
AI machines or programs capable of understanding and executing human thought patterns have been a popular theme in sci-fi for decades. Countless movies and books have explored the possibilities of future AI — and most of it looks pretty darn grim for humans — and now Google’s DeepMind firm could be bringing what was once considered fiction to life. The ushering in of new AI advancements could come at a steep price, as DeepMind trains their AI bot, named Theory of Mind-net, or ToM-net, to understand how humans think.
Indeed, the advent of a more human-like generation of AI is upon us — but how long will it take for these new machines to surpass their creators? Reports indicate that DeepMind’s new ToM-net bot is capable of passing a psychological test that “most children only develop the skills for at around age four.” It is this capacity to grasp the “theory of mind” test that may drive us towards a future of robots that are nearly as human as we are.
The new artificial intelligence out of Google-owned DeepMind is even capable of predicting what other AI bots will do. According to the DailyMail, DeepMind’s project can also understand whether or not other AIs hold “false beliefs” about their surroundings.
Has the stuff of sci-fi nightmares become reality?
As sources explain, the “theory of mind” ability is what will truly drive the more-human AI home, and ToM-net seems to have that down pat.
DeepMind engineer Neil Rabinowitz reportedly explained in a recent interview, “It can learn the differences between agents, predict how they might behave differently, and figure out when agents will have false beliefs about the world.”
“The more our machines can learn to understand others, the better they can interpret requests, help find information, explain what they’re doing, teach us new things and tailor their responses to individuals,” he continued.
While some, like the DeepMind hive, tout their new AI as a godsend, not everyone believes that highly intelligent bots are so benign. Some of the world’s most renowned minds have concerned about what more intelligent machines will spell out for humans. Elon Musk, CEO of Tesla and SpaceX, has said that AI of this caliber is the human race’s “biggest existential threat,” and also stated the development of this technology was akin to “summoning the demon.”
Musk reportedly believes that one day, super-intelligent bots might be keeping people as pets. He’s also estimated that there is a 95 percent chance that one day, our own technology advances will exterminate us.
According to the Daily Mail, late Stephen Hawking also declared it was a ‘near certainty’ humanity would be threatened by a technological “disaster” (or perhaps an uprising?) within the next 1,000 to 10,000 years. But with Google’s new tech, it seems that day could be here a lot sooner than anticipated.
It’s not just Google that we have to be worried about, unfortunately. Around the world, over-zealous scientists and engineers are working overtime to bring machines to a deeper state of understanding, without even truly considering the consequences.
One might posit that perhaps these pursuits are more ego-driven than anything else; what species actually goes out of its way to create its own successor?
With some of the world’s greatest minds cautioning against the advent of AI and warning of their dangers, you might even say these so-called experts are downright foolish. If we’ve learned anything over the years, it’s that no one ever knows as much as they think they do. Fifty years ago, smoking cigarettes was thought to be safe — and now, we know it causes cancer.
What are we going to do when AI bots decide we’re the cancer? [Related: Read more stories like this at Robotics.news]
Sources for this article include: