ChatGPT unlocks real-time info making cheating even easier

  • ChatGPT will soon be able to provide real-time information via browsing the internet.
  • Previously the popular chatbot could only access data before September 2021.
  • The removal of the date limit will make cheating at universities and schools even easier as it provided a means to catch dishonest students.

The folks at OpenAI seem bent on keeping the update train rolling for ChatGPT. The firm launched a way for the popular chatbot to verbally have conversations with users on Android and iOS earlier this week, and now another new update removes one of ChatGPT’s biggest limiters.

According to OpenAI’s X account, ChatGPT can now browse the internet and provide “current and authoritative information” complete with direct links to sources, similar to how Microsoft Bing works. This means that the free version of the chatbot is no longer bound to only provide information before September 2021 – a limit that was only removed for subscribers until now.

Before this latest update, if you were to ask ChatGPT if Donald Trump was still alive as of September 2023 it would say “I’m sorry, but I cannot provide real-time information”.

The removal of the limit now makes it even more difficult for educators to parse if students are handing in original work, or cheating with ChatGPT-made text. Lecturers at the University of the Witwatersrand, specifically in the Journalism Department, have been formulating essay questions with the date limiter of ChatGPT in mind as a way to catch cheating students.

For example, they would make sure to ask questions relating to the latest news stories or events which ChatGPT did not have access to. Now this no longer applies, or at least it won’t soon.

“Browsing is particularly useful for tasks that require up-to-date information, such as helping you with technical research, trying to choose a bike, or planning a vacation,” explains OpenAI in an X thread. It adds that browsing is now available to Plus and Enterprise subscribers but all users will receive the ability to use Browsing “soon.”

We can’t confirm if other universities in South Africa were using the same technique to limit cheating, but it is highly likely that they were using similar techniques as cheating is happening.

Daily Mail report detailed how experts estimate that more than half of college students in the UK were using the platform to cheat. Cheating has become so prevalent with ChatGPT that OpenAI itself created a tool to help educators find students abusing the AI.

The chatbot is also popular among school-age children, as evidenced by a dip in usage it saw during the summer school break in the US. As soon as the school holidays were over, usage began steadily climbing back up. While it is unknown how many primary and high school students actually use the platform for cheating, the prevalence seems high and educators are aware. Earlier in the year the New York City education department banned the ChatGPT domain on its school computers.

There is another concern that could arise out of OpenAI removing the date limit on ChatGPT’s data. Misinformation or what has been termed “hallucinations.”

Now, with the limiter off there is nothing really stopping ChatGPT from regurgitating fake news or misinformation from bad actors about politics or recent events. This could happen at a point where a prompt sees the chatbot looking for a webpage that is not necessarily true or accurate, or perhaps even completely false.

We will only come to know the ill effects of this step once the update reaches the wider free user base.

[Image – Photo by Rolf van Root on Unsplash]


About Author


Related News