ChatGPT writes well, but can give wrong answers; To avoid content with this issue, Stack Overflow has temporarily banned AI
One of the topics of the week is the Chat GPT, a chatbot that uses an AI engine to write like a human. But it didn’t take long for the initial excitement to turn into mistrust. on the platform Stack Overflowusers were prohibited from sharing responses generated by the tool.
This is a temporary ban, which can be lifted at any time. But there is a good chance that the ban will become permanent. Stack Overflow moderators study the matter to make a decision. The community can also be consulted in this process.
Why Ban ChatGPT?
You can ask ChatGPT a question or ask him to explain a concept, for example. The novelty is even capable of writing code in different programming languages. This is largely due to the foundation of ChatGPT: the GPT-3.5 AI model.
Trained with a huge amount of texts available on the web, it is GPT-3.5 which enables ChatGPT to give satisfactory explanations or convincing answers even to subjective questions. But a closer look reveals that many of these results are inaccurate or wrong.
That’s where the ban comes from. Stack Overflow is a Q&A platform for IT professionals, developers, and the like. Participants earn points when their questions or answers are voted on by the community. Based on this, many users may turn to ChatGPT to improve their reputation.
This would be a problem, as the text structure assembled by ChatGPT makes answers convincing even when they are incorrect. Stack Overflow administrators fear that the platform will be flooded with this type of content, which would make it untrustworthy.
OpenAI itself, the organization behind the initiative, warns that ChatGPT can give incorrect answers. The failure rate of the tools is not yet known. It’s just that it’s short. Despite this, the Stack Overflow admins decided not to risk it. ChatGPT has been banned from the platform, at least until the effects of its use are clear.
Right or too much decision?
Even if the ban can be lifted, the decision leaves room for several questions.
For example: Couldn’t AI increase engagement on the platform by generating more responses? Is the ban a typical attempt to stop a car from replacing a person? Wouldn’t it be the case that there was a human moderator for every ChatGPT response?
Doubts aside, I am among those who consider the temporary ban a good measure of prudence. The feature is new and its implications are not yet fully understood.
The thing is, Stack Overflow has the aspect of reliability as one of its foundations. If something can shake this hegemony, it is prudent to apply an intervention that at least allows the subject to be studied.
For now, the merit of ChatGPT is that it writes well. It may be possible to use it as a support tool, in the same way we use Google to convert a currency, for example. But, as an oracle, no. It’s still too early for that.
Stack Overflow prohibits answers provided by ChatGPT AI
🇧🇷The best content in your email for free. Choose your favorite Terra newsletter. Click here!
Source: Terra
Camila Luna is a writer at Gossipify, where she covers the latest movies and television series. With a passion for all things entertainment, Camila brings her unique perspective to her writing and offers readers an inside look at the industry. Camila is a graduate from the University of California, Los Angeles (UCLA) with a degree in English and is also a avid movie watcher.
