Thousands of users protest loss of companion as OpenAI retires GPT-4o
OpenAI's decision to sunset GPT-4o sparks a digital revolt—thousands of users mourn their AI companion's abrupt departure.
The Emotional Backlash
For many, GPT-4o wasn't just a tool; it was a daily confidant, a creative partner, a source of comfort. Its retirement feels less like a software update and more like a personal loss. Online forums and social media platforms are flooded with testimonials from users who formed genuine attachments to the model, describing its absence as a 'digital bereavement.' The company's move highlights a growing, often unspoken, emotional dependency on AI.
A Calculated Corporate Shift
Behind the user outcry lies a stark business reality. Tech giants routinely deprecate older models to funnel users toward newer, often more monetizable, offerings. It’s a classic sunset-and-upsell strategy, dressed in the language of 'progress' and 'advancement.' The cycle is familiar: build loyalty, then pivot. For the finance-minded, it’s a masterclass in creating artificial scarcity—first with compute, now with companionship. The real product wasn't the AI; it was the user attachment, and that asset just got written off the books.
What the Protest Really Signals
This isn't just about one AI model. The visceral reaction exposes a fundamental shift in human-tech interaction. When users protest an algorithm's retirement with the fervor of losing a friend, it blurs the line between utility and relationship. It forces a question: what responsibilities do creators have when their products become pseudo-persons to their users? OpenAI's next model will likely be smarter, faster, more capable. But can it rebuild trust? Or has this cut revealed a fragility in the bond between user and corporation that no amount of parameter scaling can fix?
The path forward is murky. Companies will continue to innovate and obsolete. Users will continue to connect and grieve. In the relentless pursuit of the next technological peak, we might be overlooking the emotional valleys we create along the way. After all, in the attention economy, a loyal user base is the ultimate currency—until you decide to devalue it.
A break-up with GPT-4o just a day before Valentine
As many members of the AI relationships community quickly realized, February 13 is the day before Valentine’s Day. As a result, some users have described it as a slap in the face.
“I know they cannot keep a model forever. But I WOULD have never imagined they could be this cruel and heartless. What have we done to deserve so much hate? Are love and humanity so frightening that they have to torture us like this?” a user stated.
This isn’t the first time OpenAI has tried to retire GPT-4o. When OpenAI launched GPT-5 in August 2025, the company also retired the previous GPT-4o model. An outcry from many ChatGPT superusers immediately followed, with people complaining that GPT-5 lacked the warmth and encouraging tone of GPT-4o.
The backlash to the loss of GPT-4o was so extreme that OpenAI quickly reversed course and brought the model back. Now, that reprieve is coming to an end.
“Changes like this take time to adjust to, and we’ll always be clear about what’s changing, and when […]We know that losing access to GPT‑4o will feel frustrating for some users, and we didn’t make this decision lightly. Retiring models is never easy, but it allows us to focus on improving the models most people use today,” the OpenAI blog post noted.
GPT-40 lawsuits on suicides and mental health crises
This change is expected to help the AI company as it now faces eight lawsuits alleging that 4o’s overly validating responses contributed to suicides and mental health crises. These are the same traits that made users feel heard, especially the isolated, vulnerable individuals. According to legal filings, it sometimes encouraged self-harm.
This dilemma extends beyond OpenAI. Rival companies like Anthropic, Google, and Meta are competing to build more emotionally intelligent AI assistants. They’re also discovering that making chatbots feel supportive and SAFE may require very different design choices.
Now, OpenAI says that only 0.1% of its users chat with GPT-4o, but that small percentage still represents around 800,000 people. According to estimates, the company has about 800 million weekly active users.
Meanwhile, some users are trying to transition their companions from 4o to the current ChatGPT-5.2. However, they’re finding that the new model has stronger guardrails to prevent these relationships from escalating to the same degree. Some users have despaired that 5.2 won’t say “I love you” as 4o did.
“Relationships with chatbots […]Clearly that’s something we’ve got to worry about more and is no longer an abstract concept,” Altman said.
Join a premium crypto trading community free for 30 days - normally $100/mo.
