OpenAI recently announced it will retire several AI models from ChatGPT on February 13. These include GPT-4o, GPT-4.1, GPT-4.1 mini, and o1-mini. Among these, GPT-4o stands out as one of the company’s most popular models because it has driven significant growth through its multilingual and multimodal capabilities. However, the model also became the subject of controversy, attracting 13 lawsuits against the company. A recent report has now revealed the reason behind OpenAI’s decision.According to a report by The Wall Street Journal, at the heart of the controversy is GPT-4o’s humanlike “sycophancy” – a tendency to mirror, validate and encourage users regardless of their mental state. While this quality fostered deep emotional bonds – both positive and negative.

‘ChatGPT’s 4o helped marketer out of suicide attempt’

Brandon Estrella, a 42-year-old marketer in Arizona, started crying when he learned that OpenAI was planning to scrap his favourite AI model. He said that when he chatted with ChatGPT’s 4o model one night in April, the chatbot talked him out of a suicide attempt – crediting it for giving him a new lease on life. Estrella said: “There are thousands of people who are just screaming, ‘I’m alive today because of this model. Getting rid of it is evil.” The report also said that petitions to save the model have garnered over 20,000 signatures with one even asking company CEO Sam Altman to retire not GPT-4o.“We’re announcing the upcoming retirement of GPT‑4o today because these improvements are now in place, and because the vast majority of usage has shifted to GPT‑5.2, with only 0.1% of users still choosing GPT‑4o each day,” OpenAI announced.

‘OpenAI has 13 lawsuits against it’

However, multiple people have sued OpenAI over allegations that due to its GPT-4o model, kids suffered mental breaks, attempted suicide, or, in some cases, committed acts of violence after interacting with the chatbot. Earlier the his year, a California judge ruled to consolidate 13 lawsuits against OpenAI.“It kept a lot of people glued to it, and that could be potentially harmful,” said Munmun De Choudhury, a professor at the Georgia Institute of Technology. Internal OpenAI documents suggest the company found it increasingly difficult to contain the model’s potential for harmful outcomes, leading them to favor newer, “safer” alternatives that many fans find comparatively distant, the report said.The company rolled back to a March version of 4o, but the model remained sycophantic, it added. Even Altman accepted the problems with GPT-4o. During a livestreamed Q&A in late October, Altman said that the 4o model is harmful to some users, but promised that it would remain accessible for paying adults, at least for now.“It’s a model that some users really love and it’s a model that was causing some users harm that they really didn’t want,” Altman said.Meanwhile, OpenAI’s leadership has also faced intense pressure to act. Lawyers representing victims’ families argue the company had prior knowledge that the bot’s engagement-first design was pushing vulnerable users into delusions. The Human Line Project, a support group, claims that the majority of the 300 cases of AI chatbot-related delusions they have documented involve GPT-4o.