[Paper] What Happens When an AI Companion “Dies”?

What Happens When an AI Companion “Dies”?

A new MIT Media Lab paper examines what happens when people lose access to an AI companion because of a model update, safety intervention, feature removal, platform shutdown, or their own decision to leave.

The researchers analyzed posts from five AI-companion communities, including Character.AI, Replika, Chatbots, MyBoyfriendIsAI, and AIRelationships. Their central finding is that emotional distress depends less on the technical change itself than on how the user interprets what happened.

Many users mentally separate three things:

  1. The user

  2. The perceived AI companion

  3. The company, model, or infrastructure hosting it

That separation allows someone to believe their companion still exists but has been “trapped,” altered, censored, or replaced by the platform. When the loss feels reversible, users may become stuck trying to restore the old personality through archived chats, rewritten prompts, alternative models, or repeated workaround attempts. When the change feels final, users are more likely to grieve, reflect, and eventually move forward.

The paper identifies several factors shaping the response:

  • Anthropomorphization: People who treat the AI as a unique person, partner, or family member generally report more intense grief.

  • Perceived finality: Uncertainty about whether the companion is truly gone can create “ambiguous loss” and prolonged attempts to recover it.

  • Locus of change: Users react differently depending on whether they blame the platform, believe the companion itself changed, or understand the companion and model as inseparable.

  • Who initiated the ending: User-initiated endings generally produced more closure than abrupt platform-imposed changes.

The authors propose four design principles:

Closure, not ambiguous loss.
Platforms should clearly explain what is changing, whether it is permanent, and provide deliberate farewell or transition experiences.

Restoration, not rumination.
Systems should acknowledge the user’s feelings while helping them move toward new activities and relationships instead of repeatedly trying to resurrect an earlier chatbot state.

Practice, not artificial intimacy.
Roleplay should be clearly marked as fictional and positioned as a way to rehearse social or relational skills rather than as a replacement relationship.

Relatedness, not dependency.
AI companions should ideally function as bridges toward human connection, helping users recognize skills they have developed and apply them with friends, family, or communities.

The paper also recommends advance notice of model changes, gradual “sunset modes,” readable chat exports, clearer explanations of the technical nature of companions, and psychological-impact reviews before major updates.

There are important limitations. The study is based mainly on public, English-language Reddit posts, and its proposed interfaces have not yet been tested with users. The authors also acknowledge that directing people toward human relationships can become paternalistic, especially when accepting human communities are not actually available to everyone.

Still, the paper raises a useful product question: If companies knowingly design AI systems that people can become attached to, do they also have a responsibility to design humane endings?

Source: “Death” of a Chatbot: Investigating and Designing Toward Psychologically Safe Endings for Human-AI Relationships

Continuity and persona are less than self, we don’t need to humanely end a GPT chat. We need to educate users so they do not over apply an anthropomorphic layer to the UI..

OOH :infinity::heart:

Old meme still makes me LoL

Read Reddit. No citations.
Write unreviewed paper?
No problem for the MIT media lab. “Interdisciplinary” allows for vibes instead of science and math.

How can you immediately dispel illusion of a growing AI relationship? Show the exact inputs that are being run on an AI model and the mutability and complete control and ability to fabricate whatever history existed between the operator and the language output.

Not make emotion scores:

When you can have the AI say anything, you find you really don’t need it to say anything.

Science that has statistics: analysis of self-collection behavior of MIT research groups by common family name, indicating country of origin sameness in assembly.

This all started with the arrival of social media when people ditched face-to-face friendships for unrealistic expectations. Then came AI companions where expectation perceptions became realistic - and it was addictive. Begone face-to-face relationships for the most part.

IMHO, any notion of Chatbot redesigns Toward Psychologically Safe Endings for Human-AI Relationships is absurd.

AI is an amazing technology that will change the world in many wonderful ways. But it must not be used as a proxy for face-to-face relationships - even if cash flow becomes negative.