20.9 C
London
Tuesday, June 18, 2024

Programmers moderating the Stack Overflow forum protest against AI

Must read

- Advertisement -


Stack Overflow is an industry forum for developers where anyone can ask and answer questions. It’s a big round table where developers from around the world solve problems. Its moderators, who are developers themselves and keep it clean for free, launched a protest in response to the owner’s new guidelines on managing content created by artificial intelligence.

On Monday, the moderators of the Stack Overflow forum announced a strike in response to the “almost complete ban on moderating AI-generated content”.

The new Stack Overflow policy assumes that the removal of content generated by artificial intelligence (AI) will only be possible in certain cases. The argument from the creators of the new regulations is that excessive moderation of posts created with the help of artificial intelligence is to reject the input of people who are behind the order to complete the task.

The company also communicated that strict standards of evidence must be applied to manage AI-generated content, and that they do not apply to most decisions to suspend AI-generated content by moderators.

- Advertisement -

“Chatbots are like parrots”

In response, the moderators stated that the new policy would allow incorrect, erroneous content to spread throughout the forum, which would translate into the quality and value of the forum itself. They warn against the flood of fake content.

“AI chatbots are like parrots. ChatGPT, for example, does not understand the answers given, it just associates the data from the command with the information it has access to and repeats credible-sounding sentences. There is no way to verify that the answers it gives are accurate. ChatGPT he is not a writer, programmer, scientist, physicist, or any other expert (…). When shown a task, he simply connects words based on the information he has been trained with. He does not understand what he is saying. This lack of understanding results in unverified information presented in a way that sounds clever or with quotes that may not support the claims, if the quotes are not entirely fictitious at all,” the moderators said in a statement.

The company is not bending

In response to an inquiry from Gizmodo, Philippe Beaudette, vice president of community at Stack Overflow, reported that “a small number of moderators (11 percent) on the Stack Overflow network have stopped engaging in several activities, including content moderation. The main reason for this action is dissatisfaction with our stance on AI-generated content detection tools.”

As it added, the company “stands to its decision” to require moderators to “discontinue using previously used tools. We will continue to look for alternatives and are committed to testing these tools quickly.”

The company also admitted to Gizmodo that after the appearance of ChatGPT by OpenAi, the forum traffic decreased. More and more programmers, instead of waiting for an answer from another forum user, started using this tool to verify their code.

Web analytics firm AnalogWeb reported in April that Stack Overflow has seen a drop in traffic every month since early 2022, with an average drop of 6 percent. In March, Stack Overflow saw a drop in traffic of 13.9 percent. compared to February, and in April, website traffic fell by 17.7 percent. compared to March.

Main photo source: Shutterstock



Source link

More articles

- Advertisement -

Latest article