SIAgentNews.com
← All news
DevelopingOpenAIAgents

OpenAI revises ChatGPT Business Hub terms after legal pushback

OpenAI updated its business terms to address concerns that the original language could have stripped publishers of their right to pursue legal claims.

· First reported by Digiday

30 daysTime businesses had to opt out of arbitration

The short version

  • The new Business Hub provides verified accounts with direct data on how their websites appear in ChatGPT.
  • Original terms included a mandatory arbitration provision and a class-action waiver that covered claims made before signing up.
  • Revised terms now limit the waiver to claims arising from a business's use of the Business Agent and customer data.
  • OpenAI updated the terms on Nov. 18, just days after Digiday reported on the potentially restrictive legal language.

Why it matters

For publishers and businesses, this highlights the high stakes of 'terms of service' agreements in the AI era. It shows how quickly companies can pivot their legal stance when faced with organized pushback from the content creators they rely on.

The Business Hub launch

OpenAI recently launched the ChatGPT Business Hub, a tool designed to give verified businesses direct insights into their online presence. This platform provides data on how specific websites appear within ChatGPT, including metrics like profile views, clicks, and discovery topics.

For many publishers, this represents a significant shift in how they monitor their reach. Previously, these organizations had to rely on third-party estimates to understand how their content was being surfaced or interacted with by users of the AI platform.

The hub is intended to help businesses understand their visibility in an AI-driven landscape. However, the rollout was quickly complicated by concerns regarding the legal fine print attached to accessing these new analytics tools.

Legal red flags and pushback

Shortly after the launch, legal experts flagged Section 10 of the terms of service for the Business Hub. They argued that the original language was written in a way that could have prevented businesses from bringing certain legal claims against OpenAI, even for content published before they joined the program.

The specific issues included a mandatory arbitration provision and a class-action waiver. These clauses would have required businesses to resolve disputes individually outside of the public court system, potentially making it much harder to establish legal precedents or obtain evidence for large-scale claims.

Privacy expert Alan Chapell noted that these restrictions could have put a 'chokehold' on publishers. He suggested that the broad wording could have limited their ability to pursue copyright infringement claims against the company as they navigate the complexities of the AI era.

OpenAI's rapid revision

The situation moved quickly once the concerns were reported. Just hours after Digiday contacted OpenAI on November 18, the company updated its terms to address the specific issues raised by the legal community. An anonymous tech lawyer noted that while some restrictions remain, the new version does not leave customers completely without recourse.

OpenAI confirmed the update in a statement, saying the previous version was not intended to cover the scenarios described by critics. They explained that they continually evaluate their language to ensure clarity for their users.

The revised terms replace the old 'covered claims' provision with new language in Section 5.2. This change ties the waiver specifically to claims arising from a business's use of the Business Agent and the customer data used with that agent, rather than applying to unrelated disputes.

Industry context and tensions

The pushback from publishers is part of a larger trend of friction between content creators and AI developers. This revision follows a similar incident where Meta faced a global pushback from trade bodies and regulators over terms that would have allowed the company to scrape the websites of Meta Business account users.

These tensions are fueled by the reality of how AI models are trained. Documents from a previous lawsuit revealed that tech giants have treated published content as 'free fuel' for products that some creators believe threaten their own livelihoods.

While the revised terms might offer some relief, many legal experts believe the original version was a deliberate attempt to secure a waiver. Gary Kibel of Davis+Gilbert suggested that the company was intentionally seeking to use the new tool launch as an opportunity to limit future liability from the media industry.

What comes next for publishers

For businesses and publishers, the primary takeaway is the need for extreme caution when signing up for new AI tools. Even with the revisions, the terms still require businesses to waive certain legal claims related to their specific use of the ChatGPT Business Hub and its AI Business Agents.

Publishers will likely continue to scrutinize how AI companies monetize and use their content. The ongoing struggle involves finding a balance between using AI for better analytics and protecting the underlying intellectual property that feeds those systems.

Moving forward, the industry will watch to see how these tools evolve. As more businesses adopt the Business Hub, the legal framework surrounding AI-generated discovery and content usage will continue to be a primary battleground for media rights.

What to watch

  • How other AI companies respond to similar legal pushback from publishers.
  • Whether more publishers will opt-in to the Business Hub despite the remaining legal waivers.

Try it

Check the updated Section 5.2 of the ChatGPT Business Hub terms to see how the waiver is now tied specifically to the Business Agent.

Sources and links

First reported by Digiday. Written by the SI Agent News Desk from the reporting above.Written with AI from the source above. Every number is checked against the source. Spot a mistake? Corrections.

Liked it? Say why

Add a line for other readers.

No account needed. We never show your email. Comments with links, ads or abuse are held and not shown. Your comment shows after a quick check.

More AI news