“Can AI-Powered Chatbots Truly Deliver Trustworthy News?
The Dark Side of AI-Powered Search: Can You Trust Your Answers?
In October, OpenAI launched ChatGPT search, touting a new era of search engines that combines conversational AI with the best of web search. The idea was to provide real-time information in a more useful form than just a list of links. But does this cutting-edge technology live up to its promise?
A recent review by Columbia University’s Tow Center for Digital Journalism took a closer look at ChatGPT’s performance and raised some red flags. It appears that ChatGPT’s results are prone to inaccuracies, misattributions, and even fabrications. In other words, the AI’s search capabilities can be much more problematic than initially thought.
The Problem With Accuracy
A study testing 200 quotes from 20 publications found that ChatGPT’s accuracy issues persist regardless of whether a publication blocks OpenAI’s web crawlers or has an official licensing deal with the company. Sometimes, the chatbot got it right; other times, it attributed quotes to the wrong outlet or simply made up a source. Even OpenAI’s partners, such as The Wall Street Journal, The Atlantic, and Axel Springer and Meredith publications, struggled with consistency.
Imperfect Web Crawlers, Imperfect Results
The short answer to the problem lies in the way ChatGPT finds and digests information. Even if the web crawlers are functioning properly, the AI model can still make mistakes and "hallucinate." Licensed access to content doesn’t change this fundamental fact. When a publication blocks the web crawlers, ChatGPT is left to "flounder" and fabricate sources or provide incomplete information.
The Consequences for Journalism and Trust
AI misattributing quotes is more than just a minor issue; it’s a fundamental problem for journalism and the trust we place in AI tools like ChatGPT. OpenAI wants ChatGPT search to be a reliable source for quick, accurate answers linked and cited properly. If it can’t deliver, it undermines trust in both AI and the journalism it’s summarizing. For OpenAI’s partners, the revenue from their licensing deal may not be worth the lost traffic from unreliable links and citations.
The Verdict
While ChatGPT search can be a valuable tool, it’s essential to exercise caution. When asking ChatGPT about the news, it’s crucial to verify the information and check those links to ensure the AI isn’t "hallucinating" answers from the internet.
You Might Also Like…
(Note: This section remains the same, no changes made)
