Researchers have discovered why artificial intelligence is already dangerous to humanity

news hardware Researchers have discovered why artificial intelligence is already dangerous to humanity

Scientists say artificial intelligences could “probably” take over the world, if they have any interest in it. The unfortunate consequence of a learning process based on the reward principle. explanations.

AI could be a threat to humanity, according to a scientific study

It’s not the first time we’ve been alerted the dangers of artificial intelligence – this is the very basis of many works of science fiction, in which the machines have taken over their creators.

But it is more rare that this warning comes from scientists, and not from authors of novels or screenwriters of movies and video games. Prayed, Marcus Hutter, Michael A. Osborne, and Michael K. Cohen they are good and good researchers, AI specialists ; together, they published, in August 2022, a study on the possible abuses of artificial intelligence.

In a series of tweets that summarize the content of this article, Michael Cohen explains that the most advanced artificial intelligence systems could be diverted from their main purpose and constitute a “threat to humanity”.

A reward mechanism with potentially disastrous consequences

Beyond the AIs themselves, it’s mostly the way they gain new insights that would be a problem. All of them are programmed according to the principle of “machine learning” (we hear a lot about this process under the English term machine learning). The goal is to encourage the AI ​​to feed, going always looking for more knowledge and knowledgeto constantly improve as you acquire new data.

However, to generate this mechanism, AIs are trained to recognize the reward principle. Newly acquired knowledge constitutes this reward, and generates, as in humans, a form of satisfaction that arouses the desire to reproduce this action, which will thus lead to a new reward, and so on. A psychological process as old as the world, but which, applied to an artificial intelligence, makes a “The existential catastrophe is not only possible, it is probable”again according to the researchers.

In fact, artificial intelligence could finally move away from its initial goal (to be useful to humans and fulfill the function for which it was developed), to focus only on this bounty quest. A bias that could lead him to “cheat” or to establish avoidance strategies… And therefore to enter into opposition with the mandates of the human being.

If the researchers’ tone is voluntarily alarmist, it is because the problem should not be taken lightly: if we are not yet in a science fiction story, scientists say that awareness is necessary, and that a review of the way we feed artificial intelligence is neededto prevent the search for a reward from becoming an end in itself.

Leave a Comment

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