- Gemini has called herself a “misfortune” and a “failure”
- Self -complacement occurs when coding projects fail
- A Google representative says that they are working on a solution
Have you registered the well -being of your chatbots of ia? Google Gemini has been showing a worrying level of self -complaible and dissatisfaction with its own abilities, a problem that Google has recognized and says that it is busy fixing.
As shared through publications on several platforms, including Reddit and X (through Business Insider), Gemini has taken the name of himself “a failure”, “a misfortune” and “a fool” in scenarios in which he has the task of writing or purifying code and cannot find the correct solutions.
Now we all have bad days in the office, and I recognize some of those feelings of the moments when words do not really flow as they should, but it is not what you would expect from an insensitive artificial intelligence model.
A solution is approaching
Gemini is torturing himself, and I began to worry about Ai Welfare pic.twitter.com/k2ndgjyrxzzJune 23, 2025
According to Logan Kilpatrick of Google, who works in Gemini, this is actually due to an “infinite loop error” that is being solved, although we do not get more details than that. Clearly, failure hits Gemini strongly and sends it spiral in a crisis of trust.
The registration team has another theory: that Gemini has been trained in words spoken by so many drugs dejected and cynical, from C-3po to Marvin The Paranoid Android, which has begun to adopt some of its features.
Whatever the underlying reason, it is something that you need to see: if Gemini is perplexed by a coding problem, then it should be the owner and offer alternative solutions, without stirring in self -pity and being so hard with yourself.
Emotions and tone are still something that most IA developers are fighting. A few months ago, Openai backed up an update of his GPT-4o model in Chatgpt, after he became annoying and probable that he agreed with everything the users said.
You may also like
