A controlled experiment tested four generative-AI chatbots with 200 queries about voting-related information and truth. ChatGPT Search was fully accurate on 28% of queries and completely wrong on 57%. Perplexity had a lower failure rate, being wrong 37% of the time. Additional monitoring found false claims returned to news prompts at high rates, increasing over time. Separate analysis found ChatGPT performs worst at crediting news outlets it draws from. As the 2026 midterms approach, early voters may use chatbots as their primary news interface. Research indicates the main risk is not occasional hallucination, but systematic misattribution of quotes, fabricated links, and citation practices that sever connections to original journalists.
"Chatbots misattribute quotes systematically. They fabricate links that resolve to nothing. They cite syndicated or AI-summarised copies of articles in preference to the originals, severing the chain back to the journalists. The hallucination framing is a category error inherited from the early-2024 discourse. The research shows something more specific and more dangerous for information integrity."
Read at TNW | Insider
Unable to calculate read time
Collection
[
|
...
]