M3GAN is already an icon. Could existing AI make it a reality? | CNN

Editor’s note: Reader beware, this article contains minor spoilers for “M3GAN”.


There’s no shortage of robot uprising fiction in the Western canon (see: the works of Ray Bradbury and Philip K. Dick; classics like “The Terminator”; familiar twists like “The Mitchells vs. the Machines”). It’s a conceit that plays on one of humanity’s great fears, that robots will replace us.

But there is only something about M3GAN that makes him an ideal monster and the movie in which he stars in a particularly chilling cautionary tale. (And hugely popular: “M3GAN” has already earned more than $91 million worldwide with a sequel on the way.)

True, M3GAN is an immaculately coiffed humanoid doll with killer instincts and a distinctive fashion sense. It seems unlikely that an advanced and, frankly, cheeky robot will be taking over the world any time soon. But his eponymous film was also released during the advent of advanced AI that can hold conversations and write essays or create intricate works of digital art.

“M3GAN is a metaphor for many things that happen in our lives, (including) the unintended consequences of autonomous robotics,” Daniel H. Wilson, a science fiction author and former robotics specialist, said in an interview with CNN.

“M3GAN” raises questions about our human tendency to trust technology, including artificial intelligence, which is already seamlessly integrated into many of our lives, even if it doesn’t take the form of a well-dressed robot tucking kids to bed. . We regularly consult Siri or Alexa, the integrated assistants of Apple and Amazon products, respectively; many consumers clamor to buy Teslas, with its semi-automatic capabilities; and most of us can spend hours scrolling through social media thanks to algorithms meant to maximize the amount of time and attention we spend on those platforms.

Much of “M3GAN” is pure fiction – there are currently no robots that can move as fluidly as M3GAN can when charging at enemies at full speed. But the AI ​​model driving M3GAN is real, even if it’s not immediately available, said Shelly Palmer, a resident professor of advanced media at Syracuse University and an expert in fields including emerging technology.

“The technology is absolutely coming, but the real wild card is human beings,” Wilson said.

Mechanically, no, the technology hasn’t caught up to do M3GAN in real life, both Palmer and Wilson said. It’s not yet possible to make a bipedal robot that runs and turns fluidly and perfectly executes TikTok dances before executing humans she perceives as threats.

You may be familiar with the work of Boston Dynamics, the company whose four-legged robots can walk with relative ease and carry heavy objects (and which inspired a man-versus-machine episode of “Black Mirror”). Wilson and Palmer pointed to the company as a leader in the field of robotics, with fully mobile bots already being used in industries including construction and law enforcement (the company says Spot is used to test potentially toxic materials or hazardous environments). He hasn’t even created a robot comparable to M3GAN.

But the AI ​​systems that power M3GAN in the film are already in the works, Palmer and WIlson said. In a blog post, Palmer wrote that “AGI systems,” or “artificial general intelligence,” which describes a bot’s ability to learn everything a human can learn, will soon be ready to roll out.

“The technology for a wrist like M3GAN is absolutely here, and an affordable version is coming at some point,” said “Robopocalypse” author Wilson. “The question is whether we want it.” (Based on humanity’s well-documented fears of dolls and a robot-dominated future, many of us probably won’t.)

Wilson’s “How to Survive a Robot Uprising” was published in 2005 and technology has evolved tremendously since then. At the time, he said that he did not believe a robot uprising was inevitable. but there is is It’s technology that worries him today, he told CNN, and it’s not robots that take on a physical form, but AI products that mimic the reactions of humans.

“Mostly we are afraid that robots will replace us,” Wilson said. “For a long time we were afraid that jobs were going to replace us… but now I think they are going to replace us in intimate places, in relationships.” (Some may remember this being a key theme in the 2013 film, “Ella.”)

Pointed to M3GAN: She acts as a friend to young Cady, who loses her parents in a car accident. M3GAN fills the comforting role given by Gemma, a cold robotic struggling to adequately comfort her orphaned niece. But advanced AI doesn’t expect reciprocity. Humans may not look to other humans for friends, those weakness-prone people, because they can get whatever they want from robots, which are designed to serve, Wilson said.

Gemma, played by Allison Williams (right), gifted her grieving niece an M3GAN doll to help meet the girl's emotional needs.

AI already hurts people “every day,” Palmer said, albeit in less visible ways than, say, M3GAN slamming a young bully into oncoming traffic. Social media addiction, she said, is driven by AI, as is the “prolific spread of misinformation.

“The sci-fi version of robots or AI falling into a ‘man vs. machine’ conflict is embellished to create good visuals; the truth is much more subtle and has been going on for years,” Palmer said.

New tools, like the popular chatbot ChatGPT, can faithfully reproduce human speech, but as with any form of social networking, you also want to keep users’ attention, Wilson said.

“What scares me is that social networks are combined with generative artificial networks: ChatGPT could lead a whole generation of people into a conspiracy theory,” he said. “Social media just wants to command your attention and will say or do anything they can to maximize those interactions.”

ChatGPT and other forms of AI aren’t inherently bad, Palmer said, but they will almost certainly be used in ways their creators didn’t intend. If there is a widely available form of technology, be it a rock, a gun, or a cell phone, someone will try to misuse it.

“Most technology is neither good nor bad,” Palmer said. “That distinction is reserved for individuals.”

Wilson put himself in the level-headed shoes of Allison Williams’ characters to explain how a robotics specialist could avoid turning a hypothetical M3GAN into a killing machine.

“If I wanted to make M3GAN a product and make sure it didn’t have a deadly robotic rampage, I would integrate some design features that make the robot unable to kill people,” he said.

He would make his M3GAN physically weak so that it would not be able to throw humans or heavy objects. He would not implement a black box model, which prevents users from understanding how an AI product works, making it easier to run diagnostics on the robotic wrist. But the most important thing, he said, is to ensure that M3GAN is hacker-proof. If a malevolent hacker can infiltrate M3GAN, he could use it to access private information or cause physical damage, he said.

And on the off chance that M3GAN does in fact go rogue, Wilson said some advice from his first book still stands: “Always look for the sensors first.” Take out the cameras that your robotic enemy uses to see you and you have already won.

Leave a Reply

Your email address will not be published. Required fields are marked *