Moderators of Stack Overflow, a Q&A platform for programmers, go on strike to protest the company's allowing AI-generated content to flood the site

News on June 7, Stack Overflow is a well-known software developer forum operated by Stack Exchange, where users can ask and answer various programming questions.

The moderators wrote in an open letter: "Stack Overflow, Inc. has banned any censorship of AI-generated content... allowed by default on the Stack Exchange network The emergence of incorrect information ('hallucinations') and unbridled plagiarism. This poses a significant threat to the integrity and credibility of the platform and its content. In protest of this and a number of other recent and upcoming policy and platform changes, we The decision has been made to launch an immediate, full-scale moderator strike on the Stack Overflow and Stack Exchange networks."

Moderators can no longer censor content because it is generated by artificial intelligence, this policy will be implemented at the end of May . However, moderators believe that without proper curation of AI-generated content, the quality and accuracy of information on the Stack Exchange site will rapidly degrade.

The moderators published an article on Meta Stack Exchange further explaining their appeal. The article reads, "AI chatbots are like parrots. ChatGPT, for example, doesn't understand the answers it gives you; it just retells sentences that seem reasonable based on the prompts it's given and the information it has access to. It can't verify it. that the answers provided to you are accurate. ChatGPT is not a writer, a programmer, a scientist, a physicist, or any other type of expert upon which our website relies. When asked, it is based solely on its training to piece words together. It doesn’t understand what it’s saying.”

The article also states, “This would allow users to paraphrase ChatGPT’s answers without understanding what they are saying, which defeats the purpose of our website: Become a repository of high-quality Q&A content."

Moderators say another problem is the company's lack of transparency in developing this policy. In December, the site implemented a temporary policy banning all use of ChatGPT because it was "generally inaccurate" and violated the site's citation requirements, they wrote in the article. This policy was supported by volunteer moderators and Stack Exchange staff and resulted in many posts being removed and users being banned.

However, on May 29, moderators wrote that the company had privately implemented a new policy requiring that it “immediately cease banning users for AI-generated content and based solely on this moderating AI-generated content for reasons. The next day, a slightly different version of the policy was published, which did not include language asking moderators to stop restricting all AI content.

"The new policy overturned The consensus that the community has reached, without discussion with any community members, is misleading to both moderators and the public, and is based on unfounded claims based on unvetted and unreviewable data analysis," Meta The Stack Exchange article reads.

The moderators also asked the company to improve non-AI related issues on the site. "This strike is also largely about the recent behavior of the Stack Exchange company. "This behavior pattern," the article reads, "the company once again ignored the needs of the community and the existing consensus, and instead focused on business transformation at the expense of its own community managers." They cited the example of a site's chat feature that was outdated but had been neglected for years.

IT House noted that the strike was the first major action against the influx of AI-generated content into online sites, and moderators on other forums had similar concerns. Some Reddit moderators have braced themselves for a number of AI-generated posts with misinformation. An unnamed Reddit moderator told Motherboard that the ChatGPT-powered bot problem on Reddit is "pretty serious" and that hundreds of accounts have been manually removed from the site because Reddit's automated systems struggle to cope with AI-generated bots. content.

The moderators asked the company to retract and revise its AI policy to resolve the inconsistency between the public and private versions, and apologize to the public. Philippe Beaudette, vice president of community at Stack Overflow, told Motherboard in a statement, "A small number of moderators (11%) on the Stack Overflow network have stopped participating in some activities, including moderating content. The main reason for our action is due to concerns about our Dissatisfied with the position on AI content generation detection tools. Stack Overflow conducted an analysis and found that the ChatGPT detection tool previously used by moderators had an alarmingly high false positive rate." The moderators wrote in the article that they knew the detection tool There is a problem with the tool. "We stand by our decision to ask moderators to stop using the previous tool," Beaudette continued, "and we believe we will find a path forward. We regret that things have reached this stage, and the community management team is evaluating the situation while Try to stabilize the situation in a short time."

The above is the detailed content of Moderators of Stack Overflow, a Q&A platform for programmers, go on strike to protest the company's allowing AI-generated content to flood the site. For more information, please follow other related articles on the PHP Chinese website!