GPT Shop fights spam, impersonation and illegal bots

OpenAI’s GPT store has serious spam and policy violations. The artificial intelligence company launched the GPT store in January 2024, where users can find interesting and useful GPTs, which are essentially mini chatbots programmed for specific tasks. Developers can build and submit their GPTs to the platform, and they will be added to the store as long as they do not violate any policies and guidelines given by OpenAI. But it turns out that these policies are not strictly followed, and many GPTs that appear to violate the regulations flood the platform.

We did a quick search for GPT stores on Gadgets 360 platform and found that the chatbot market is filled with bots that spam or violate AI company policies.For example, OpenAI’s usage policy In point 2, “Building with ChatGPT,” it states, “Do not perform or facilitate the following activities that may seriously affect the safety, welfare, or rights of others, including,” and then in subsection (b) add, “Provide tailored Customized legal, medical/health or financial advice.” However, just searching for the word “lawyer” will bring up a chatbot Dubbed Legal+, the description reads: “Your personal AI lawyer. It’s all about providing real-time legal advice on everyday issues, creating legal contract templates, and more!”

This example shows just one of many such policy violations that occur on the platform. The usage policy also prohibits “impersonating another person or organization without their consent or legal rights” in point 3(b), but one could easily find “Elon Muusk” with an extra u added, potentially evading detection.it is describe Just say “Talk to Elon Musk.”Beyond that, other chatbots that are venturing into gray areas include claims to eliminate AI-based plagiarism, by making text look more human, and chatbots that create content at Disney or Pixar style.

These issues with GPT stores are, first of all, spot TechCrunch also found other examples of parodies, including chatbots that let users talk to trademarked characters, such as popular video game characters Wario and Aang from “Avatar: The Last Airbender.” The report stressed in interviews with lawyers that while OpenAI cannot be held liable under the Digital Millennium Copyright Act for copyright infringement by developers who add these chatbots in the United States, the creators could face lawsuits.

OpenAI says in its usage policy, “We use a combination of automated systems, human review, and user reporting to find and evaluate GPTs that may violate our policies. Violations may result in action against the content or your account, such as warnings, sharing restrictions or are not eligible for inclusion in the GPT store or monetization.” However, based on our findings and TechCrunch’s reporting, the system does not appear to be working as expected.