AI barking is worse than being bitten.
As fear-mongering about the chances of AI reaches high levels, experts say so-called advanced technology is not anywhere near as smart as your dog.
Thread!
At Wednesday’s VivaTech conference, Yann LeCun, AI Meta’s chief scientist, revealed that currently generative AI is “still very limited.”
“These systems are still very limited. They haven’t any understanding of the essential reality of the actual world because they’re only trained in text, huge amounts of text,” said LeCun, for CNBC.
In contrast, much of what people learn “has nothing to do with language” and at this stage of technological progress “that a part of the human experience shouldn’t be captured by AI.”
In other words, robots – while they could excel in medical and business exams – are unable to learn basic duties corresponding to unloading a dishwasher.
![Yann LeCun](https://nypost.com/wp-content/uploads/sites/2/2023/06/NYPICHPDPICT000012679157.jpg?w=1024)
![Golden Lab with a tennis ball in the park](https://nypost.com/wp-content/uploads/sites/2/2023/06/NYPICHPDPICT000012749737.jpg?w=768)
“What does it let you know [is] we’re missing something really big… To attain not only human-level intelligence, but even dog-level intelligence,” LeCun said.
In truth, experts “don’t know easy methods to replicate that performance with today’s machines.”
“Until we are in a position to try this, we won’t have human-level intelligence, we won’t have dog-level intelligence or cat-level intelligence [intelligence]LeCun added.
But his remarks contrast with those of Google DeepMind CEO Demis Hassabis, who said that AI could achieve “human-level cognitive abilities” inside the subsequent decade.
Hassabis’s announcement comes just days after Geoffrey Hinton, often known as the Godfather of AI, quit Google to precise his concerns and regrets concerning the development of technology, while people, mostly employees, fear for his or her jobs, and others worry about security.
![Open the robot's brain with wires and a motherboard inside](https://nypost.com/wp-content/uploads/sites/2/2023/06/NYPICHPDPICT000010923389.jpg?w=1024)
![Vann LeCun](https://nypost.com/wp-content/uploads/sites/2/2023/06/NYPICHPDPICT000012679670.jpg?w=1024)
“I console myself with the conventional excuse: if I hadn’t done it, another person would have,” the 75-year-old previously said.
He isn’t the just one – Sam Altman, CEO of OpenAI, he was vocal concerning the potential threats posed by artificial intelligence.
ChatGPT, a sensible chatbot software created by OpenAI, has been used to cheat by college students and dubbed the “atomic bomb” by business tycoon Warren Buffett.
But as users tinker with the smart bot, they realize that is not all it’s.
In accordance with the showrunners of the hit Netflix series “Black Mirror”, ChatGPT was not fun as entertainment.
![A robot with the NASA logo](https://nypost.com/wp-content/uploads/sites/2/2023/06/NYPICHPDPICT000011357723-1.jpg?w=1024)
Charlie Brooker, the creator of the series, said that the chatbot wrote the script of the tantalizing drama.
“Because all that is been done is looking up all of the ‘Black Mirror’ episode summaries and kind of mixing them together,” Brooker said this month. “Then for those who dig somewhat deeper, you are like, ‘Oh, there’s actually no real original thought here.’