Ai-Da the robot and the myth of AI



This week, a robot “gave evidence” in the House of Lords – an event that is being celebrated by the robot’s developer and the media as a groundbreaking historic moment.

The robot, called Ai-Da, is billed as “the world’s first ultra-realistic robot”. It’s the brainchild of artistic entrepreneur Aidan Meller, a gallery director with no technical training. Engineering students from the University of Leeds have designed a “drawing arm”, allowing the dummy to generate primitive doodles (or “art” as the media called it). Another team created a silicone face and added a wig. This was eventually put together by Engineered Arts, a company founded by special effects expert Will Jackson, which creates animatronic mannequins.

The result looks suspiciously like Sky News Westminster correspondent Beth Rigby. And just like a real human, Ai-Da “can also lean back and forth, wave his arms, and look around in all directions,” making him a match for any detective in the SW1 lobby. .

Once his robot had a body, Meller began to advertise. He describes Ai-Da as “breathtaking” and “groundbreaking” – claims the enthusiastic media took at face value. Ai-Da and her exhibits have been featured in over 900 publications. Some media even “interviewed” Ai-Da. Questions must be pre-submitted to Meller in advance. Then he generates the responses so that they can be ventriloquized by Ai-Da.

Last year, in a stunt worthy of late Sex Pistols manager Malcolm McLaren, Ai-Da was “arrested” in Egypt for espionage. Meller told the press that “the British ambassador was working all night” to publish it in time for an art exhibition. Luckily for everyone, that deadline was met.

This week, Ai-Da and Meller were summoned by the House of Lords Communications and Digital Committee, as part of its inquiry into “the future of the UK’s creative industries”. As with Ai-Da’s interviews with the media, the questions and answers were prepared in advance.

During the session, Meller sat next to the dummy as he read her responses, most of which offered little more than book transhumanism. “I don’t have subjective experiences… [but] I always create art,” Ai-Da told the Lords.

Despite the hype, Ai-Da is not the first robot to get the call from parliamentarians seeking publicity. Four years ago, Pepper, a talking dummy created by Middlesex University engineering students, ‘gave evidence’ to a committee set up to discuss the ‘fourth industrial revolution’. MPs speculated that Pepper’s successors would play a big role in the economy for years to come.

“The Fourth Industrial Revolution is perhaps the most significant challenge our nation will face in the next 10, 20 to 30 years,” committee chair Robert Halfon said in 2018. The Pepper session, like the one of Ai-Da, also took the form of pre-arranged questions and scripted answers. The two sessions could just as well have been an exchange of emails. Unlike Pinnochio, neither Pepper nor Ai-Da developed an agency – they just ventriloquized the thoughts of their creators. The physical robot is superfluous. In 2018, Wired‘s Gian M Volpicelli called the Pepper session “an unnecessary parliamentary pantomime”.

This week’s House of Lords stunt deserves the same contempt. Or maybe even more. Since at one point, Ai-Da couldn’t even finish her canned responses and had to be restarted.

Last month, Tesla CEO Elon Musk also revealed a humanoid robot, Optimus. Musk says Optimus could one day do dangerous manual labor, help out in nursing homes, or become a sexual partner. The media – usually more than willing to accept outlandish claims about robotics – were more skeptical on this occasion. Robotics experts were given the opportunity to report the defects. Robots can “perform simple tasks and exhibit superficial levels of situational awareness, but all it takes is one detail to be off the mark and the robot will quickly descend into a series of errors that usually end in an irreparable disaster”, notes an artificial intelligence researcher. Filip Piekniewski in Business Intern. According to Piekniewski, Optimus is a “total and utter scam”. Even pro-Tesla critics admit that Optimus didn’t really do much to advance the field of robotics.

Five years ago, speaking to parliamentarians about the future of AI and robotics, I urged them to take evidence from more skeptical experts – including Piekniewski, Gary Marcus and Jaron Lanier – to get a more realistic in the field. None have ever been summoned. I concluded that parliamentarians care much more about being seen to be ahead of the technology curve than about the technology itself. This week’s Ai-Da session only confirmed this suspicion.

The truth is, there hasn’t been a quantum leap in robotics for at least a decade. There have been steady advancements in robot capabilities and deployment. But there have also been setbacks.

The most egregious failures have come from attempts to attach “artificial intelligence” to machines. Amazon recently halted the expansion of its human-free robot stores in the UK and also cut Prime robot deliveries. Level four or level five autonomous driving – which would mean high or full automation – is receding further and further into the future. Robotics is progressing, but there is no sign of a “revolution”.

Computer scientist Jaron Lanier warned 20 years ago that “we make ourselves stupid in order to make computer software intelligent”, recalling that “artificial intelligence is best understood as a belief system rather than a technology”. This now sounds like a chilling and prescient description of today’s political class. They are so eager to stand at the bow of the ship of state, their eyes glued to the horizon as it sails into a bold technological future, that they feel compelled to create the entirely imaginary world of ” fourth industrial revolution”. And if that means they have to stoop in front of an animated mannequin, that’s part of the job.

Source link


Comments are closed.