Barely integrated into Bing, ChatGPT is already being abused by a privileged few who have obtained access to the functionality.
In a veritable tidal wave of registrations only a few days after its deployment, the very young chatbot was thus pushed to its limits, to the point of making incoherent remarks and exposing its errors.
Barely arrived on Bing, ChatGPT is hazed
One thing is certain, Microsoft managed to shine the spotlight on its AI-powered Bing browser last week. Curious users responded in large numbers… for better or for worse.
Some lucky ones have indeed had fun, so to speak, “hazing” the newly arrived chatbot in Bing, which has sometimes literally derailed. Several such examples have been posted on the Bing subreddit, such as this conversation where the AI goes into an existential crisis after losing the thread of the conversation.
Another classic when dealing with AI: a user asked Bing’s chatbot if he was conscious. The answer is reminiscent of a reaction that we have already seen at least once in a work of science fiction featuring artificial intelligences.
The mistakes of youth
In addition to the malicious pleasure taken by some users in breaking up the very recent Bing chatbot, others wanted to highlight the errors it is capable of generating. To prove that Bing’s new AI is untrustworthy, Dmitri Brereton notably posted on his blog, cited in source below, some examples of rather gross errors.
For the time being, this seems to lack coherence in the responses it generates. She can even transcribe in an erroneous way information yet marked black on white on documents that she has found to respond to a directive.
A Microsoft representative spoke about these childhood mistakes: We expect the system to make mistakes during its testing period, and feedback is critically important to help us identify what is not working in order to make our model better. »
The Bing chatbot being still, so to speak, a “baby”, it is far from being a threat to the search engines that we are used to, contrary to what one might fear.
Download
- Chat in different languages, including French
- Generate, translate and get a text summary
- Generate, optimize and correct code
Developed by OpenAI, ChatGPT is a conversational agent based on the GPT-3 language model. This chat platform uses deep learning and artificial intelligence to understand and interpret user queries, then generate relevant responses.
Developed by OpenAI, ChatGPT is a conversational agent based on the GPT-3 language model. This chat platform uses deep learning and artificial intelligence to understand and interpret user queries, then generate relevant responses.
Sources: Reddit, DKB-Blog
15