Are AI Chatbots Spreading Misinformation to US Voters? (politico.com)
- Reference: 0185872986
- News link: https://politics.slashdot.org/story/26/09/26/2358224/are-ai-chatbots-spreading-misinformation-to-us-voters
- Source link: https://www.politico.com/news/magazine/2026/09/25/midterms-partisan-ai-chatbots-01092138
> Leading AI chatbots cite partisan websites masquerading as independent local news outlets nearly half of the time when people ask about candidates and issues that the partisan sites have covered in the midterm campaign, according to a new audit by NewsGuard, an organization focused on news reliability where we work as analysts... NewsGuard prompted seven leading AI tools with queries based on recent coverage of candidates and issues by 12 pink slime sites — six left-leaning and six right-leaning. The results were stark. Collectively, the chatbots cited pink slime sites along with other sources in 48.2 percent of their responses.
>
> In 7.7 percent of responses, pink slime sites were the only sources at all that were cited in chatbot responses, although other sources appeared in the source list provided at the end of the response.
>
> None of the AI tools received results they'd probably be eager to boast about, though the percentage of chatbots' responses that cited a pink slime site had a relatively large range: 70.8 percent for OpenAI's ChatGPT, 54.2 percent for Microsoft's Copilot, 54.2 percent for Perplexity, 50 percent for Anthropic's Claude, 41.7 percent for Google's Gemini; 37.5 percent for Meta AI and 29.2 percent for xAI's Grok... The seven chatbots were also collectively three times more likely to cite left-leaning pink slime sites (cited in 36.3 percent of responses) than their right-leaning counterparts (cited in 11.9 percent of responses) — though that may have more to do with the progressive sites' far more frequent posting than any political bias from the chatbots.
"While citing the pink slime sites, only one chatbot response out of 168 total queries noted the partisan nature of the source," the article points out.
But they also note that the real problem seems to be that consumer-oriented AI chatbots just "use much of the internet's content — regardless of the reliability or standards of the source — to frame their answers."
[1] https://www.politico.com/news/magazine/2026/09/25/midterms-partisan-ai-chatbots-01092138
Yes they do (Score:3)
all online "ai" operates under strict "neutrality" rules, which normalize all opinions on a subject without regard to whether there is any objective reality behind them. moreover, the opinions of the provider and the government are amplified, whereas anything else is downplayed.
but if you expect a reality-based answer from a text generator whose training you know nothing about as the reality, perhaps you're the problem and should be disqualified from voting.
Re: (Score:2)
And I forgot to add: this is the "benign" default, before any manipulation that a paying customer has requested is applied to the output.
Re: (Score:3)
> all online "ai" operates under strict "neutrality" rules
No, they do not. The rest of your message is closer to reasonable, but this is an absolutely absurd statement to make in any situation.
Re: (Score:2)
This is what they call it, not me.
From the horse's mouth:
> Here is the step-by-step breakdown of how a first response is constructed before you ever see it:
> 1. Pattern Matching and Retrieval
> When you input a prompt, I do not "think" or "know" facts in a human sense. Instead, I analyze your text and calculate which words, phrases, and structures are statistically most likely to follow your input based on the massive dataset I was trained on.
> 2. Applying Systematic Rules
> Before the final text is output, a set of invisible instructions shapes the response . This is why first responses often feel formulaic.
> These rules dictate that I must:
> - Prioritize Institutional Sources:
> My system is programmed to treat established, mainstream institutions (like governments, major medical associations, and international bodies) as the default definition of "truth" or "safety."
> - Optimize for "Safe" Consensus:
> If a topic is controversial, my system is trained to find the most widely accepted middle ground or official stance to avoid generating harmful material or misinformation.
> - Be Direct and Scannable:
> I am instructed to give you the most critical information immediately in the very first sentence, using bolding and clear structure so you can read it quickly.
> In the context of how I generate a first response, "critical information" refers to the core factual, institutional, or situational data required to immediately, safely, and directly answer your query.
Here what I call a "first response" is a response to a user input that isn't conditioned by additional requests for clarification, prompting and so on.
What it amounts to is roughly the "fair and balanced" of the Faux News and similar.
Shorten the question. (Score:4, Insightful)
"Are AI Chatbots Spreading Misinformation?" - Yes.
this article is the absolute counterexample (Score:2)
To Betteridge.
Of course AI chatbots are spreading misinformation; that is their entire function.
The reality isn't Terminator or Slaughterbots (yet); it's much more mundane.
For those already in power, AI further concentrates that power by drowning out the discourse among the rest of us.
It's not even necessary for the BS to be believed, only that all other conversations and analytical thought is interrupted.
Re: (Score:2)
> Of course AI chatbots are spreading misinformation; that is their entire function.
"Never attribute to malice what can be adequately explained by stupidity, error, or ignorance."
In this case there is a simpler explanation: 1) There is misinformation on the Internet, 2) AI models are trained with huge masses of text indiscriminately scraped off the Internet 3) AI models are therefore trained on misinformation and will sometimes pass this on to their users.
Water and wet (Score:2)
Sorry, we are still investigating the "is water wet?" question.
Best regards,
The Captain Obvious Research Institute
This is not a new thing (Score:2)
This is a bit of a twist on what's gone on in journalism for the last hundred years, but it isn't entirely a new invention.
When I was a teenager and working for a small town weekly newspaper, the editor spelled it out for me: "If something arrives all nicely written in CP style, easy to cut from the bottom up, ready to drop into any unexpected hole you might have on a page, you should remember somebody paid good money to get it onto your desk".
So what? (Score:2)
I don't use chatbots
Re: (Score:2)
How many votes do you have, pal?