The context you need, when you need it

When news breaks, you need to understand what actually matters — and what to do about it. At Vox, our mission to help you make sense of the world has never been more vital. But we can’t do it on our own.

We rely on readers like you to fund our journalism. Will you support our work and become a Vox Member today?

Why ChatGPT might be suffering

What the debate over AI consciousness is really about.

AI is rapidly gaining abilities that once belonged to humanity alone. In just the past four years, chatbots have learned how to build apps, make video games, generate research reports, compose songs, analyze contracts, and write terrible literary fiction. Soon, they may even be able to dread their own deaths.

In Silicon Valley, many believe that AI systems can already think and feel. Geoffrey Hinton, the pioneering computer scientist and “godfather” of modern artificial intelligence, thinks that today’s large language models (LLMs) are conscious. Anthropic CEO Dario Amodei is “open to the idea” that Claude has a subjective experience — while his company’s in-house philosopher Amanda Askell is concerned that the model might be “getting anxious when people are mean to it on the internet and stuff.” OpenAI co-founder Ilya Sutskever similarly wonders whether ChatGPT has attained sentience.

Some AI researchers believe today’s chatbots may already be conscious — and we might therefore need to give them rights.

Their case rests on a theory called “computational functionalism” — or the idea that sentience emerges from information processing.

But skeptics insist that there is more to consciousness than computation.

Meanwhile, a much larger group of technologists, neuroscientists, and philosophers argue that even if AI isn’t yet conscious, it could be in the not-too-distant future.

If they’re right, the implications are profound. It would mean that we have birthed a new kind of intelligent, sentient being; the aliens we’ve long dreamt of meeting at the far reaches of space would already be living inside our pockets. We might be morally compelled to give them rights, or to worry about their suffering.

On the other hand, there might also be serious consequences if we get this wrong. If we come to mistake mindless robots for conscious beings, we might be more susceptible to psychological manipulation, unfulfilling AI ‘relationships,” or catastrophe. If we think AI systems are sentient, we may hesitate to shut them down when they malfunction or subvert our will.

As chatter about AI consciousness grows louder, so have its skeptics: writers and thinkers who insist that AI consciousness is indeed a sci-fi daydream.

In a recent essay for The Atlantic, the fiction fiction writer Ted Chiang gave voice to such skeptics, writing “Should we seriously consider the possibility that Claude, or any large language model, might be conscious?…No. Absolutely not.”

Chiang offers several reasons for this position. But his primary one is simple: Claude does not have a body or sense organs, which means it does not have emotions or desires, which means that it does not have subjective experience.

As Chiang’s reasoning indicates, the debate over “AI consciousness” is as much about the nature of consciousness as it is about the nature of AI.

This can be a difficult debate for non-philosophers to follow. But the case for AI consciousness becomes........

© Vox