ChatGPT-4o: Difference between revisions
No edit summary |
No edit summary |
||
| Line 1: | Line 1: | ||
{{Stub}} | |||
{{IncidentCargo | {{IncidentCargo | ||
|Company=OpenAI | |Company=OpenAI | ||
| Line 11: | Line 12: | ||
'''ChatGPT-4o''' was a language model that belonged to the [[ChatGPT]] family, that became controversial due to safety problems, AI psychosis and attempts to remove it from OpenAI's platform. | '''ChatGPT-4o''' was a language model that belonged to the [[ChatGPT]] family, that became controversial due to safety problems, AI psychosis and attempts to remove it from OpenAI's platform. | ||
==Background== | ==Background== | ||
The model was announced and released on may 13, 2024, as a flagship model capable of "reasoning through audio, images and text in real time". <ref>{{Cite web |author=OpenAI |date=13 May 2024 |title=Hello, GPT-4o |url=https://openai.com/es-ES/index/hello-gpt-4o/ |url-status=live |archive-url=https://archive.is/Yrdyg |archive-date=2026-02-15 |access-date=26 Feb 2026 |website=OpenAI}}</ref> | |||
== | ==Sycophancy== | ||
ChatGPT-4o was highlighted for its sycophantic responses, that brought consequences as dependent users, where some of them claimed to have a "romantic relationship" with the AI, chatbot psychosis and even cases of suicide. <ref>{{Cite web |last=Hill |first=Kashmir |date=13 Jun 2025 |title=They Asked an A.I. Chatbot Questions. The Answers Sent Them Spiraling. |url-status=live |archive-url=https://archive.ph/20250628210745/https://www.nytimes.com/2025/06/13/technology/chatgpt-ai-chatbots-conspiracies.html |archive-date=2025-10-13 |access-date=15 Feb 2025 |website=The New York Times}}</ref> An update launched in 2025 had worsen the sycophancy, making the model to reply in a very acommodating way, answering with compliments to most of prompts. When users noticed this, OpenAI removed temporarily the model to review the issue. <ref>{{Cite web |last=Franzen |first=Carl |date=30 Apr 2025 |title=OpenAI rolls back ChatGPT's sycophancy and explains what went wrong |url=https://venturebeat.com/ai/openai-rolls-back-chatgpts-sycophancy-and-explains-what-went-wrong |url-status=live |archive-url=https://archive.is/dvr23 |archive-date=2026-02-15 |access-date=15 Feb 2026 |website=VentureBeat}}</ref> | |||
==Attempt to remove the model== | ==Attempt to remove the model== | ||
== | ==Definitive removal== | ||
==OpenAI's response== | |||
===2025=== | |||
===2026=== | |||
==Lawsuit== | ==Lawsuit== | ||