12:23 GMT06 March 2021
Listen Live
    Get short URL
    0 21

    The dangers of the unregulated development of artificial intelligence and AI-powered robots have been previously cited by many experts, including SpaceX CEO Elon Musk, who warned about the difficulty in predicting their behaviour once unleashed.

    Author of a newly released book called “The AI Does not Want to Kill You” Tom Chivers told Daily Star Online that a conversation with the world’s most renowned artificial intelligence experts has revealed that the creation of “superhuman AI” will likely occur in the next fifty years, and could potentially lead to either immortality or the complete destruction of human civilisation.

    Chivers cited the conversation with Google Android engineer Paul Crowley to illustrate that experts were currently divided on the issue of what humanity will face with the rapid advancement of powerful AI machines. The author of the groundbreaking book was said to be particularly “horrified” by Crowley’s remarks. 

    “His reasoning was if you look at what most AI researchers think about when really powerful superhuman AI will come, most of them think it will happen in roughly the next 50 to 100 years”, said Chivers as quoted by the media.

    “And then also if you ask them [AI experts] what the outcome of something like this will be, building some incredibly powerful machine like that, a good chunk of them reckon that there is a 20-30% chance it will be human civilisation-ending. It literally will kill everyone”.

    He noted that many specialists believe the creation of superhuman AI will help to end poverty and cure diseases, thus leading to immortality.

    “It is the idea that we could be on the verge of something really amazing or completely catastrophic, within my kids’ lifetimes and possibly in my one”, Chivers added. “I found that really hard to take on board because it is inherently weird and we sort-of assume the future will be basically like today”.

    In the book, which is now available online, Chivers also cited an experiment where several AI robots were instructed to play tic-tac-toe, with one winning by ultimately “forcing its opponents to crash” via cheating algorithms.

    Several very prominent tech experts, including Tesla and SpaceX CEO Elon Musk, have previously expressed their concerns over the unregulated development of artificial intelligence and growing capabilities of AI-powered robots. It remains unclear whether superhuman AI machines will be able to follow human instructions in the future or whether their behaviour will be completely unpredictable. 

    robots, Elon Musk, artificial intelligence
    Community standardsDiscussion