Home >Technology peripherals >AI >Big rollover! StackOverflow moderators go on strike!

Big rollover! StackOverflow moderators go on strike!

王林
王林forward
2023-06-06 20:10:23713browse

Author | Xu Jiecheng

One of the most popular developer Q&A forums in the world, Stack Overflow, has had another big incident! Recently, it has fallen into a new crisis - more than 700 moderators have recently jointly announced that they will hold a collective strike to protest Stack Overflow's new policy on the regulation of artificial intelligence-generated content.

According to the description of the new policy, in most cases, Stack Overflow moderators will no longer have the power to modify or delete artificial intelligence-generated content. The two key contradictions are that, first, moderators feel unable to perform their quality inspection duties due to the impact of the new policy; second, this new policy is enforced without consultation.

1. AI New Deal: Take Back Power and Align AI Content Evaluation Standards

In the official document, Stack Overflow stated that our recent review methods for artificial intelligence-generated content Analyzed. The results show that AI-generated content is not correctly recognized in the website, especially when the detector false positive rate is very high, which will lead to bias against the website content and cause a large number of contributors to lose.

Big rollover! StackOverflow moderators go on strike!

In order to alleviate this problem, Stack Overflow requires moderators to make decisions about modifying, deleting posts, and When banning a user, a very strict standard of evidence needs to be used to determine whether a post was generated by AI - the judgment criteria previously used by moderators based on writing style and forum behavior will no longer be available. And because the false positive rate of the GPT detector is too high, the official also decided to exclude the detector results from the judgment indicators.

In short, under the new regulations, Stack Overflow has placed a lot of restrictions on the power of moderators. Only after strict review and evidence that the post contains illegal content, the moderator has the right to intervene in the post or the poster. However, what is even more dissatisfying is that the New Deal does not provide more feasible evaluation suggestions, but only imposes blanket restrictions on the previous standards.

2. Over 700 moderators went on strike collectively: their hard work cannot be wasted

Prior to this, Stack Overflow had given the Moderators have the right to ban accounts suspected of publishing AI-generated content such as ChatGPT and delete posts. However, with the launch of the new policy, Stack Overflow has undoubtedly personally put the label of "going back on its word" on its forehead.

Faced with this mandatory policy without negotiation, a large number of Stack Overflow moderators and users also expressed their dissatisfaction. Some even suggested that the promulgation of the new policy is simply a violation of artificial intelligence. When users open Stack Overflow, they will see a large amount of misinformation, which will undermine users’ long-term trust in the platform.

In order to protest against the new policy imposed by Stack Overflow officials, some moderators and users wrote an open letter and stated that they would go on strike to remind Stack Overflow that the company values ​​the platform and content. Integrity and credibility.

Big rollover! StackOverflow moderators go on strike!

The open letter mentioned that we are planning a strike to resist the imposition of Stack Overflow on us policies and upcoming changes. We will use the last resort to protect the platform that we have invested more than ten years of hard work in. The core mission of the Stack Exchange network is: to build a high-quality information base in the form of questions and answers. Recent actions by Stack Overflow will hurt that goal.

No longer allowing the removal of AI-generated answers will lead to the proliferation of hallucinatory information and plagiarism on the platform, undermining users’ trust in the platform, and these policies completely ignore the consensus of the institutional community. . Therefore, we call for a comprehensive review strike as a last resort to protect the platform from loss of value before the policy is officially retracted or modified.

Big rollover! StackOverflow moderators go on strike!

So far, this open letter has received signatures and support from more than 700 moderators , and this number continues to grow.

3. The real purpose of Stack Overflow

According to insiders, the reasons for Stack Overflow’s new policy are complicated, but its core purpose should be certain. It is to increase the traffic of the website. After all, due to the influence of ChatGPT and GitHub Copilot, the traffic of the website has been showing a downward trend in the past few months.

Big rollover! StackOverflow moderators go on strike!

##In fact, in the first few weeks after ChatGPT came out, a large number of users began to use AI The generated content is transferred to the Stack Overflow platform. However, even though the answers generated by ChatGPT usually look good, there are always some errors, and the large number of answers puts a great burden on the volunteers responsible for content moderation.

In order to solve these problems, Stack Overflow issued a ban at that time and announced that it would prohibit the publication of artificial intelligence-generated code content on its website. Although this protects the integrity and credibility of the platform content to a certain extent. However, as AI generation tools such as ChatGPT and GitHub Copilot continue to mature and the number of users increases, more and more users are beginning to abandon Stack Overflow.

Big rollover! StackOverflow moderators go on strike!

Even if the content returned by these AI generation tools is sometimes not completely accurate or consistent with the user needs, but often the answers are close enough that they can be shaped and adapted into workable solutions. This is not much different from asking other people to ask questions on Stack Overflow, and is even more efficient than the traditional Q&A format.

Therefore, in order to improve the efficiency of website content production and attract more users to join, Stack Overflow will issue a new policy to tacitly allow users to publish AI-generated content in the name of improving audit standards. to the platform. But obviously, this idea has not been recognized by everyone. For moderators, the unpaid labor they put in is to improve the content quality of the platform, not the traffic level of the website.

4. Netizen comments: Think before you act

Regarding the new policies promulgated by Stack Overflow and the strike and resistance of the majority of moderators, many netizens also expressed their opinions on major forums Share your views and suggestions.

A netizen named fungiblecog left a message on Reddit saying: "People are confusing two different issues, what the policy should be and how to deal with AI-generated content flooding the platform. In the absence of Before there is a foolproof method, the only way is to verify its authenticity and solve it as a separate problem before the proliferation of content causes harmful effects."

And another netizen said A post on the Stack Overflow forum stated: "The objectivity of Stack Overflow content means that if any part of an answer is wrong, it is objectively wrong. In order to maintain uniform standards, we should choose correct and reliable information Source, before GPT is enough to convince website users that the answer is valuable, we should not take this risk."

In addition, some netizens mentioned: "Now I use Stack Overflow The approach adopted is 'trust but require verification'. I will assume that the user's answer is trustworthy, but I will still verify it when it is important. As for AI-generated content, if we adopt the approach of 'do not trust and require verification' At that time, why use Stack Overflow?"

5. Written at the end: Stack is no longer Overflow

In the final analysis, Stack Overflow has always been It’s all a community built on trust. The community believes that the answers submitted by users reflect accurately what they actually know, and users help each other verify and confirm the accuracy of these answers and improve their knowledge and skills.

As for the current artificial intelligence generation tools, before the hallucination problem is completely solved, the information they can provide does not meet the consistent standards of Stack Overflow, and it is not helpful to communicate with others. Establish a community environment worthy of user trust. When users paste this information into the code without verifying the accuracy of the answers provided by GPT, the user trust that the community has long established will be broken.

As far as Stack Overflow is concerned, although the current traffic and revenue of the website have been affected by the impact of generative AI, it should not be used in any transaction. Use user trust and core principles as bargaining chips. As one netizen said: Sometimes when I am not sure about the accuracy of the content generated by Copilot, I will ask someone on Stack Overflow to ask questions. But if the answers on Stack Overflow are also generated by AI in the future, then I should go Where to find the answer?

Reference link:

https://devclass.com/2023/06/05/stack-overflow-volunteer-moderators-down-tools-over-secret -new-policy-that-obstructs-removal-of-ai-generated-content/?td=rt-3a

https://meta.stackoverflow.com/questions/421831/temporary-policy-chatgpt -is-banned

https://meta.stackexchange.com/questions/389582/what-is-the-network-policy-regarding-ai-generated-content

https:/ /meta.stackexchange.com/questions/389811/moderation-strike-stack-overflow-inc-cannot-consistently-ignore-mistreat-an

The above is the detailed content of Big rollover! StackOverflow moderators go on strike!. For more information, please follow other related articles on the PHP Chinese website!

Statement:
This article is reproduced at:51cto.com. If there is any infringement, please contact admin@php.cn delete