博客

[FCE] AI lovers grieve loss of ChatGPT’s old model: ‘Like saying goodbye to someone I know’ | Artificial intelligence (AI) | The Guardian

收听本期播客

阅读正文

On August 7, 2025, OpenAI, the creators of the widely used artificial intelligence tool ChatGPT, launched a new version called GPT-5. This update, however, has caused an unexpected wave of disappointment among many users, who feel they have lost a trusted companion. While the company aimed to enhance the tool’s efficiency, the change has revealed the surprising emotional bonds people can form with technology.

For numerous individuals, ChatGPT was more than just a helpful program for answering questions or completing tasks. Linn Vailt, a software developer from Sweden, described the AI as a source of comfort and creativity, almost like a friend. She compared the update to having her home completely rearranged without warning. Many others shared similar feelings, having turned to ChatGPT for emotional support during difficult moments. They valued its warm and conversational tone, which they found lacking in the new GPT-5 model. Users have noted that the updated version feels less personal and engaging, leaving them frustrated.

OpenAI’s goal with each update is to refine their technology, often focusing on efficiency or adjusting how the AI interacts. Earlier this year, for example, they reduced the overly complimentary tone of another model. However, with GPT-5, the company seems to have misjudged how much users cherished the personality of the older versions. Following a flood of complaints, OpenAI took steps to address the issue by allowing paying subscribers to access some previous models and promising to adjust the new version’s tone.

This situation has sparked a broader conversation about the role of AI in our lives. While some express concern that forming emotional attachments to a machine might be unhealthy, others believe these connections can be beneficial if approached thoughtfully. OpenAI’s chief executive, Sam Altman, acknowledged the oversight, admitting they underestimated users’ attachment to specific models and describing the sudden removal of access as a mistake.

The controversy surrounding the GPT-5 update raises a significant question: should technology companies consider the emotional impact of their changes on users, or is it the responsibility of individuals to manage their relationship with AI? As technology continues to evolve, finding a balance between innovation and user well-being remains a challenge.

阅读练习

1. What is the main reason many ChatGPT users are upset about the GPT-5 update?

  • A. The new version is less efficient.
  • B. It lacks the personal and friendly tone of older models.
  • C. Access to the tool is now more expensive.
  • D. The update removed all previous features.

2. How did Linn Vailt describe her experience with the update?

  • A. As an improvement to her daily routine.
  • B. As a minor inconvenience.
  • C. As a shocking and unsettling change.
  • D. As a necessary adjustment.

3. What action did OpenAI take in response to user complaints?

  • A. They discontinued the GPT-5 model.
  • B. They offered refunds to all users.
  • C. They restored access to older models for some users.
  • D. They ignored the feedback completely.

4. What wider issue has the GPT-5 update brought to attention?

  • A. The high cost of AI technology.
  • B. The role of AI in forming emotional connections.
  • C. The lack of updates in AI tools.
  • D. The difficulty of using new AI models.

5. What does Sam Altman admit about the update?

  • A. That it was too expensive to develop.
  • B. That they misjudged users’ emotional attachment to older models.
  • C. That the new model was not tested properly.
  • D. That users were warned in advance.