META CEO Mark Zuckerberg has unveiled a ruthless, more lifelike model of artificial intelligence.
Named after the Roman statesman, the model, named Cicero, is capable of performing human skills such as negotiation, trickery and foresight, according to the Washington Post reported.
In addition, the AI algorithm often wins in the complex strategy game Diplomacy.
This makes Cicero the first AI computer to achieve human performance in the complex strategy game with natural language.
“An agent that can play at the human level in a game as strategically complex as diplomacy is a real breakthrough for cooperative AI,” said Yann LeCun, VP and lead AI scientist at Meta AI.
The AI played with humans on webDiplomacy.net, the online version of the game.
Playing more than 40 games, Cicero displayed deceit, ruthlessness, and cunning.
Additionally, after playing with real people, Cicero ranked in the top 10 percent of participants who played more than one game.
Meta chose diplomacy because the AI has been evolving for decades thanks to the standards set by the gameplay.
Cicero’s success in Diplomacy proves just how impressive – and lifelike – AI models can become, with Meta calling it a “breakthrough” in a blog post.
The tech giant added that the breakthrough lies in combining two different areas of AI: strategic thinking and natural language processing.
“The integration of these techniques gives Cicero the ability to reason and strategize in light of player motivations, then use natural language to communicate, make agreements to achieve common goals, form alliances, and coordinate plans ‘ said Meta.
However, some experts worry that AI has the ability to withhold information or think ahead.
“This type of technology could be used to concoct smarter scams that blackmail people, or create more convincing deepfakes,” writes Pranshu Verma for the Washington Post.
“This is a great example of how much we can fool other people,” Kentaro Toyama, a professor and expert in artificial intelligence at the University of Michigan, told the Post.
“These things are super scary… [and] could be used for evil.”
“AI is like the nuclear power of this age,” Toyama added.
“It has tremendous potential for both good and bad, but… I think if we don’t start regulating the bad, all the dystopian AI sci-fi will become dystopian science fact.”
https://www.thesun.ie/tech/news-tech/9834522/mark-zuckerberg-ruthless-power-hungry-ai-life/ Mark Zuckerberg creates a “ruthless and power-seeking” AI that is eerily lifelike