ChatGPT and AI could redefine how we think about consciousness

ChatGPT and AI could redefine how we think about consciousness

ChatGPT and other new chatbots are so good at mimicking human interaction that they have prompted a question among some: Is there any chance they are aware?

The answer, at least for now, is no. Almost everyone who works in the field of artificial technology is sure that ChatGPT is not alive in the way it is generally understood by the average person.

But that is not where the question ends. What it means to be conscious in the age of artificial intelligence is up for debate.

“These deep neural networks, these matrices of millions of numbers, how do you map that to these views that we have about what consciousness is? That’s a kind of terra incognita,” said Nick Bostrom, founding director of the Future of Humanity Institute at Oxford University, using the Latin term for “unknown territory.”

The creation of artificial life has been the subject of science fiction for decades, while philosophers have spent decades considering the nature of consciousness. Some people have even argued that some AI programs, as they now exist, should be considered sentient (a Google engineer was fired for making such a claim).

Ilya Sutskever, co-founder of OpenAI, the company behind ChatGPT, has speculated that the algorithms behind your company’s creations might be “faintly aware.”

NBC News spoke to five people who study the concept of consciousness about whether an advanced chatbot could possess some degree of consciousness. And if so, what moral obligations does humanity have towards such a creature?

It is a relatively new area of ​​research.

“This is a very recent area of ​​research,” Bostrom said. “There’s a lot of work that hasn’t been done.”

In true philosophical fashion, the experts said that it really comes down to how you define the terms and the question.

ChatGPT, along with similar programs like Sydney, Microsoft’s search assistant, are already being used to help with tasks like programming and typing simple text like Press releases, thanks to its ease of use and convincing command of English and other languages. They are often referred to as “extensive language models” as their fluency comes largely from having been trained on gigantic troves of text culled from the Internet. While his words are compelling, they are not precisely crafted as a top priority and are notoriously frequent. mistaken when they try to state facts.

Spokespersons for ChatGPT and Microsoft told NBC News that they follow strict ethical guidelines, but did not provide details on concerns that their products could raise awareness. A Microsoft spokesperson stressed that the Bing chatbot “cannot think or learn on its own.”

In a long post on your websiteStephen Wolfram, a computer scientist, noted that ChatGPT and other large language models use mathematics to calculate the probability of which word to use in a given context based on the text library it was trained on.

Many philosophers agree that for something to be conscious, it has to have a subjective experience. in the classic paper “What is it like to be a bat?” The philosopher Thomas Nagel argued that something is conscious only “if there is something that is like be that organism.” A bat is likely to have some kind of bat-like experience, even though its brain and senses are very different from a human’s. A dinner plate, by contrast, would not.

David Chalmers, co-director of the Center for Mind, Brain and Consciousness at New York University, He has written that while ChatGPT clearly does not possess many of the commonly assumed elements of consciousness, such as sensation and independent agency, it is easy to imagine that a more sophisticated program could.

“They are excellent liars.”

Susan Schneider, founding director of the Center for the Future Mind at Florida Atlantic University.

“They are like chameleons: they can adopt any new character at any time. It is not clear that they have goals and core beliefs that drive their action,” Chalmers told NBC News. But over time they can develop a clearer sense of agency, he said.

One problem philosophers point to is that users can go ahead and ask a sophisticated chatbot for insider experiences, but they can’t trust it to give a reliable answer.

“They are excellent liars,” said Susan Schneider, founding director of Florida Atlantic University’s Center for Future Minds.

“They are increasingly capable of having more and more fluid interactions with humans,” he said. “They can tell you that they feel that they are people. And then 10 minutes later, in a different conversation, they’ll say the opposite.”

Schneider has noted that current chatbots use existing human writing to describe their internal state. So one way to test whether a program is conscious, he argues, is to not give it access to that kind of material and see if it can still describe the subjective experience.

“Ask him if he understands the idea of ​​survival after the death of his system. Or if he would miss a human he interacts with often. And you investigate the responses and find out why he reports the way he does,” she said.

Robert Long, philosophy fellow at the Center for AI Safety, a San Francisco nonprofit, cautioned that just because a system like ChatGPT has complexity doesn’t mean it has conscience. But on the other hand, he pointed out that just because a chatbot can’t be trusted to describe its own subjective experience doesn’t mean it doesn’t have one.

“If a parrot says ‘I am in pain,’ this does not mean that it is in pain, but it is very likely that parrots are in pain,” Long wrote in his substack.

Long also said in an interview with NBC News that human consciousness is an evolutionary byproduct, which could be a lesson in how an increasingly complex artificial intelligence system could come close to a human idea of ​​subjective experience.

Something similar could happen with artificial intelligence, Long said.

“Maybe you don’t intend to, but through your effort to build more complex machines, you might get some kind of convergence in the kind of mind that has conscious experiences,” he said.

The idea that humans could create another type of sentient being raises the question of whether they have any moral obligation to it. Bostrom said that while it was hard to speculate on something so theoretical, humans could start by simply asking an AI what it wanted and agreeing to help with the easiest requests: “low-hanging fruit.”

That might even mean changing your code.

“It may not be practical to give it all at once. I mean, I would like to have a billion dollars,” Bostrom said. “But if there are really trivial things that we could give them, like just changing one little thing in the code, that could matter a lot. If someone has to rewrite a line in the code and suddenly they’re much happier with their situation, then maybe they’ll do it.”

If humanity eventually ends up sharing the earth with a synthetic consciousness, that could force societies to drastically reevaluate some things.

Most free societies agree that people should be free to reproduce if they choose, and also that a person should be able to have a vote for representative political leadership. But that gets thorny with computer intelligence, Bostrom said.

“If you’re an AI that could make a million copies of itself over the course of 20 minutes, and then each of them has a vote, then something has to give,” he said.

“Some of these principles that we think are really fundamental and important should be rethought in the context of a world where we cohabit with digital minds,” Bostrom said.

By Mitchell G. Patton

You May Also Like