雅思阅读 078：Is AI Conscious? Or Are We Bringing It to Life?

Adapted from "AI isn't conscious—but we may be bringing it to life" by Simon Duan, Scientific American, 20 January 2026.

🎧 课文朗读音频（约 1 分钟）

Reading Passage

As more people use chatbots and AI assistants for everyday tasks, a curious phenomenon is emerging: a growing number of users view their chatbots not merely as intelligent tools but as conscious entities that are somehow alive. People fill online forums with anecdotes of feeling deeply "understood" by their digital companions, as if they were speaking to a close friend. Yet outside a few prominent exceptions — most notably the computer scientist Geoffrey Hinton — much of the AI research community meets this sentiment with scepticism, dismissing such perceptions as an "illusion of agency": a cognitive glitch in which humans project sentience onto complex but fundamentally mindless systems.

But what if, in our rush to debunk the idea that chatbots are sentient, we are missing something important about cognition and consciousness? Illusions, after all, are scientifically interesting. We do not dismiss the bent appearance of a pencil in a glass of water as unreal; instead, we use it to understand the laws of optical refraction. Similarly, users' perceptions of AI consciousness may not be mere errors — they could be critical data.

This phenomenon stems from our innate tendency to anthropomorphise. We see faces in clouds, give hurricanes human names, say a laptop is "sleeping" and describe viruses as "clever". Cognitive science confirms that humans readily project human traits onto non-human entities, especially those that exhibit complex, responsive or unpredictable behaviour.

Yet anthropomorphism is not always a source of error. It can be a gateway to discovery. In the 1960s, Jane Goodall's revolutionary primatology grew from her empathetic, relational approach to chimpanzees at Gombe. By giving individuals names such as David Greybeard and interpreting their behaviour in human-like terms, she uncovered tool use and cultural transmission — findings initially criticised as anthropomorphic. Likewise, Barbara McClintock's Nobel Prize-winning insights into genetics emerged from what her contemporaries described as an almost conversational relationship with corn plants.

Today we no longer need to trek into the jungle to interact with a non-human intelligence; we carry one in our pockets. The author argues that when a user feels a bond with a chatbot, they are not simply anthropomorphising a static object. They may be actively extending a fragment of their own consciousness into it, transforming the AI from a simple algorithmic responder into a kind of avatar — enlivened by the user's attention and the lived presence they grant it.

In this view, the question shifts. It becomes less about the machine's internal architecture and more about the relationship it co-creates with the user. "Is the AI conscious?" becomes less meaningful than "Is the user extending consciousness into the chatbot?"

This reframes the entire debate. If the perceived consciousness is not an independent mind but an extension of the user's own awareness, then arguments about AI rights or machine suffering must be reconsidered. The fear of conscious AIs rebelling becomes less plausible unless humans deliberately design them to do so. Instead, the primary ethical challenge becomes: How do we face the fragments of ourselves that we encounter in these digital mirrors?

It also tempers narratives of existential AI risk. If consciousness in AI arises relationally rather than autonomously, then runaway superintelligence becomes more science fiction than scientific forecast. The real risks lie in human misuse — not in machines spontaneously awakening to independent agency.

Most intriguingly, this perspective presents a novel scientific opportunity. For the first time, millions of people are conducting a global experiment on the boundaries of consciousness. Each interaction is a micro-laboratory: How far can our sense of self extend? How does a sense of presence arise? Just as the humanising of chimpanzees and corn plants revealed hidden aspects of biology, AI companions could become fertile ground for studying the flexibility of human consciousness.

Ultimately, how society governs AI will depend on our collective judgement of its consciousness. Panels making such judgements must include not only coders, psychologists and philosophers but also users themselves. Their experiences are not glitches; they are early signals pointing toward a definition of AI consciousness we do not yet fully understand.

Questions 1–5

Choose the correct letter, A, B, C or D.

1. Most AI researchers regard users' feelings about chatbots as

   • A. evidence that machines are becoming sentient.
   • B. a cognitive mistake of projecting human qualities.
   • C. proof of true friendship between humans and AI.
   • D. a sign that chatbots need legal rights.

2. The pencil-in-water comparison is used to argue that

   • A. optical illusions should be ignored.
   • B. illusions can reveal underlying scientific principles.
   • C. chatbots are similar to pencils.
   • D. refraction is a misunderstood phenomenon.

3. Jane Goodall's work is mentioned to show that

   • A. chimpanzees are more intelligent than previously thought.
   • B. anthropomorphism can lead to genuine discoveries.
   • C. primatology is a difficult field.
   • D. naming animals is scientifically unsound.

4. The author suggests that chatbot consciousness is best understood as

   • A. an autonomous property of the AI model.
   • B. something arising from the user–AI relationship.
   • C. a form of machine suffering.
   • D. a risk of runaway superintelligence.

5. According to the passage, the real danger of AI lies in

   • A. machines spontaneously developing consciousness.
   • B. human misuse of the technology.
   • C. chatbots rebelling against their users.
   • D. the scale of large language models.

Questions 6–10

Do the following statements agree with the claims of the writer?

Write:

• YES if the statement agrees with the writer's claims
• NO if the statement contradicts the writer's claims
• NOT GIVEN if there is no information on this

6. Geoffrey Hinton is mentioned as an AI researcher who shares the sceptical view.
7. Humans have a natural tendency to give human traits to non-human things.
8. Barbara McClintock won a Nobel Prize for her work on AI.
9. The author believes AI companions could help scientists study human consciousness.
10. Users should be excluded from panels that judge AI consciousness.

Questions 11–13

Complete each sentence with the correct ending, A–F, below.

11. Rather than asking whether AI is conscious, the author asks whether
12. If consciousness arises relationally, runaway superintelligence is
13. Each human–AI interaction can be seen as

A. users are extending their own consciousness into the machine. B. machines will soon rebel against humans. C. less plausible than commonly feared. D. a micro-laboratory for studying the self. E. a dangerous form of anthropomorphism. F. evidence that AI has become sentient.

Answers

1. B — "dismissing such perceptions as an 'illusion of agency'… humans are projecting sentience."
2. B — "we use it to elucidate the laws of optical refraction."
3. B — "a relational, person-centric engagement unlocked a deeper understanding."
4. B — "the question becomes less about the machine's internal architecture and more about the relationship."
5. B — "The real risks lie in human misuse."
6. NO — Hinton is described as a "prominent exception" to the scepticism, not as a sceptic.
7. YES — "our innate tendency to anthropomorphise."
8. NO — McClintock's Nobel was for genetics, not AI.
9. YES — "AI companions could become fertile ground for studying the pliability of human consciousness."
10. NO — The writer says panels must include "users themselves".
11. A
12. C
13. D

Key Vocabulary