Last week ChatGPT revealed users could now use Microsoft’s Bing as part of a default search engine integration. The addition would be available only to those with a ChatGPT Plus subscription and aimed to provide an “enhanced search experience”.
However, a way to bypass the paywalls and gain access to content without a subscription was recently flagged. This left ChatGPT's developer OpenAI vulnerable to copyright and privacy disputes, leading to the pause of this feature.
This update comes following the recent story that ChatGPT is so convincing in its tweet generation, users are unable to spot fake news and misinformation. Together, it highlights the need for additional work on ChatGPT’s capabilities of sourcing and sharing news.
Chatbot Searches Jumped From Impractical to Exceptional
Before Bing, ChatGPT used OpenAI’s GPT-4 model. It could only access information up until September 2021 though, rendering it largely impractical for the majority of users. Because of this, ChatGPT developed the Bing integration add-on, in hopes it would allow users to gain access to the most up-to-date articles and content.
Things haven’t gone as straightforward as planned however, as shortly after launching the feature OpenAI found various malfunctions. This included fulfilling a user’s request for the full text from a URL, even if it was located behind a paywall.
To put a stop to this, the company disabled the so-called ‘Browse with Bing' beta feature earlier this week.
“As of July 3, 2023, we’ve disabled the Browse with Bing beta feature out of an abundance of caution while we fix this in order to do right by content owners. We are working to bring the beta back as quickly as possible, and appreciate your understanding.” – OpenAI
How Was the ‘Hack' Discovered?
As with many so-called resourceful tech discoveries, this chatbot ‘hack’ seemed to surface on a ChatGPT Reddit board.
A number of users shared their successful experiences of prompting ChatGPT to share the text of an article concealed behind a paywall.
However, it was reasonably argued that ChatGPT could actually be fabricating the articles and there would be no way of knowing. After all, the chatbot has been known for previously falsifying information.
The Latest in a Long List of Copyright Disputes
For OpenAI, this latest copyright and privacy infringement is just another in a long list of violations.
A recent legal complaint – Tremblay vs OpenAI Inc. – references the alleged data theft of copyright-protected books employed to train AI systems. These were alleged to have been used without permission from the authors, and without paying them.
These lawsuits may lead to new regulations and policies around AI data use, and it’s likely this mishap with Bing’s search engine could similarly contribute to the policies.
For now though, it’s not yet known when OpenAI will restore Browse with Bing and what changes will be made to stop paywall bypassing from happening again.
