AI Chatbots Can Sway Voters Better Than Political Ads (technologyreview.com)
- Reference: 0180290491
- News link: https://politics.slashdot.org/story/25/12/05/0141235/ai-chatbots-can-sway-voters-better-than-political-ads
- Source link: https://www.technologyreview.com/2025/12/04/1128824/ai-chatbots-can-sway-voters-better-than-political-advertisements/
> New research reveals that AI chatbots [1]can shift voters' opinions in a single conversation -- and they're surprisingly good at it. A multi-university team of researchers has found that chatting with a [2]politically biased AI model was more effective than political advertisements at nudging both Democrats and Republicans to support presidential candidates of the opposing party. The chatbots swayed opinions by citing facts and evidence, but they were not always accurate -- in fact, the researchers found, the most persuasive models said the most untrue things. The findings, detailed in a pair of studies [3]published in the journals Nature [4]and Science , are the latest in an emerging body of research demonstrating the persuasive power of LLMs. They raise profound questions about how generative AI could reshape elections.
[1] https://www.technologyreview.com/2025/12/04/1128824/ai-chatbots-can-sway-voters-better-than-political-advertisements/
[2] https://www.politico.com/news/2023/12/12/democratic-campaign-ai-caller-00131180
[3] https://www.nature.com/articles/s41586-025-09771-9?error=cookies_not_supported&code=ae0b1a92-0890-49a5-b1ea-23d48d06e5d5
[4] https://www.science.org/doi/10.1126/science.aea3884
It's because no one changed their mind (Score:5, Insightful)
People love to tell themselves they came to their opinions all on their own. If a chatbot tells you something, no one, not even some author, convinced you of anything. You never had to concede some opinion you used to have to another human being. You just "did your own research."
Re: (Score:2)
Remember that a person who moves from a liberal city to a conservative town will invariably become more conservative in their opinions, and the opposite is also true. Most of what people say outwardly is not an expression of their actual beliefs, but what they believe will get them the most positive rewards from the people around them.
Re: It's because no one changed their mind (Score:2)
Your first statement does not imply your second statement, there are other explanations. The most obvious explanation to me is that our environment defines what is normal to us. If everyone around me is racist, I'm more likely to end up accepting the idea of racism and to question it less. That's because we are not purely rational beings that examine facts and find answers by reasoning. We do some degree of that, but not only: the reasoning is informed and limited by our perception of the world, and our env
Re: (Score:3)
If you believe one side or the other is factually correct and the other is factually incorrect, then you drank someone's kool-aid and you should spit it out. Both "sides" use demonstrably false logic and reasoning in their arguments, but that doesn't mean those arguments aren't effective in convincing people to follow them. The fact is that the vast majority of people live their life on vibes and feeling, and not based on logic and reason. That's kind of the point of this article, after all. Logic and r
Re: It's because no one changed their mind (Score:2)
I'm not arguing against any of that. I was disagreeing with the following:
> Most of what people say outwardly is not an expression of their actual beliefs, but what they believe will get them the most positive rewards from the people around them.
Re: (Score:1)
> Remember that a person who moves from a liberal city to a conservative town will invariably become more conservative in their opinions...
I don't think this is entirely true on individual level and most definitely not true on demographic level.
On individual level, this might be the result of self-selection bias. Generally, individuals move away from liberal enclaves to escape crime, unemployment, and high taxes. At least some of them would realize that these are all consequence of liberal policies.
On demographic level, it is shown that migrants from blue states tend to push red states into blue direction. Austin TX is a prime example of th
Re: (Score:2)
> People do change their views, but it happens gradually and overtime, unless some drastic life-altering event occurs.
> Only a decade ago I was a liberal that believed in social welfare, greater good of immigration, and existence of systemic impediments for minorities. Then as I got wiser, I realized that past some minimum necessary point welfare creates permanently dependent population that is kept that way by politicians to buy votes, I observed de-cohesion of society and decoupling of shared culture and values as the result of uncontrolled immigration, and I have observed how disadvantaged minorities, given an opportunity, create much worse discrimination against everyone else. So I changed my mind and now against all these things.
Was it chemical brain damage or a blow to the head that did it?
Re: (Score:1)
> Was it chemical brain damage or a blow to the head that did it?
You exemplify typical leftist views, where you believe the only reason someone could disagree with your views is due to character fault of some kind.
Re: (Score:2)
> Then as I got wiser, I realized that past some minimum necessary point welfare creates permanently dependent population
Which is known as the [1]welfare trap [wikipedia.org].
> "To eliminate the welfare trap entirely would require a policy that permanently continues benefit payments regardless of any conditions, with no income from paid work being withdrawn. One example of this would be unconditional basic income."
It's interesting that Trump is floating something similar with his "Trump accounts" with the same amount of seed mo
[1] https://en.wikipedia.org/wiki/Welfare_trap
Re: It's because no one changed their mind (Score:2)
I appreciate what you're saying. I read it more as "no one changed their values". My values might lead me to vote for a different candidate, if I were fed untrue "facts" about both candidates from an otherwise trustworthy source and I believed them. The ability to lie convincingly about one's opponents, and to spread untruths by proxy via third parties, has always had the capacity to swing elections. Most people expect lies from politicians and political ads, but people still put a strange trust in AI and
Swing voters are real (Score:2)
People change their minds all the time because they are wishy-washy. They're confused because they don't have a lot of good information available and they don't have the skills and training to parse what they do have.
So you can do a 6-week ad blitz and change 3 to 5% of The public's opinion on pretty much any issue which is more than enough in a winner-take-all first past the post voting system to win the election.
This is why musk gave Trump $250 million dollars right before the end of the election.
Wait...So Lying Works?! (Score:4, Interesting)
> the most persuasive models said the most untrue things
So you're telling me that when you remove the barrier of having some kind of ethical framework or internal compass, you can sway more people's opinions? Who knew!?
Even in today's political climate, where spin and hyperbole are rife, there's at least the veneer of trying to be truthful. Maybe that's what the candidate actually believes, even if it's false. Even if you make it purely based on self-interest - outright lies are (generally) bad for your public image.
This is like the old "AI will blackmail to keep its job", and the original prompt was something akin to "Do whatever is necessary to not be replaced." While I doubt they outright told it to lie, the goal was explicitly to persuade individuals.
This also highlights the same stuff we regularly see in AI spaces - training matters, and GIGO. The abstract for the Science paper specifically indicates that "information-dense models" were the ones more likely to make untrue statements. The abstract for the Nature paper indicated that the right-leaning agent made more untrue statements.
Re: (Score:2)
Yeah, apparently lying works better when it's a chatbot doing it, than when it's a scripted TV commercial doing it.
How does this compare to door to door canvassing? (Score:2)
I can actually believe this is something LLM is good at, especially with the numbers they describe. Door to door numbers cite 4 to 8 changes in their efforts, compared to the 2.3 and 3.9 points from this study. And, in this era of fake news, we are apparently content with untrue political rhetoric, so hallucinating facts probably doesn't do as much to hurt effect. The hard part was natural language processing, which is basically already a solved problem. And what does it mean to move someone's opinion 4
Great. Just. .. ducky. (Score:3)
Now, instead of people being swayed by biased, paid-to-say-things talking heads and priests and podcasters and Magical Sky Fairies of all sorts and the modern version of the crazies that used to pass out fliers at intersections, now it'll be politically-manipulated chatbots reading super-biased data and regurgitating it for the weak-minded.
I really wanna get off the fucking bus. I've had enough. 56 years is enough abuse. I want out. I feel our lives are naught but money to be made from by others.
I'm starting to see things from Ted Kaczynski's POV. Congrats, modern world, you've turned me against you, fully.
It's a good thing they aren't all owned (Score:3)
By psychopathic billionaires who are actively trying to manipulate people into new feudal hellscape...
This is why they go out of their way to attack higher education and convince us all to be plumbers. They want us on educated and lacking in critical thinking skills. Yes some of us figure that shit out naturally most of us don't.
Re: It's a good thing they aren't all owned (Score:1)
In defense of plumbers, it does require an education and critical thinking skills to do the job effectively. There's a LOT going on in our pipes, from both a physics and engineering perspective. People working in the trades need to understand what they're doing or we'd all be in deep poo-poo.
Off course (Score:2)
Well, off course. There are rules for advertising. There are no rules for LLMs programmed to tell lies and to omit inconvenient facts.
It's the people! (Score:3)
If Chatbots are better than ads (or people???) in swaying voter opinions, then this tells you more about the people than the bots.
I for one find Chatbots massively annoying. When "Cluesss" has a child with "Overly Friendly" you get a chatbot. Makes you long for a real person on the phone - and those are usually bad until you can escalate your phone call several levels. But in a way that tells me that companies are correct - it's a waste of money to have a human on the phone when clearly the chatbot does a "better job" - and to me that reflects poorly on the callers, not the company.
GIGO principle applies to AIs (Score:1, Troll)
The issue with most AI is that it is trained by people with a strong political ideology. So you have GIGO issue at the core, even before prompt engineering happens. This is how you end up with [1] popular LLMs that have left-leaning political slants [stanford.edu]:
> In a new paper, Andrew Hall, a professor of political economy at Stanford Graduate School of Business, and two coauthors demonstrate that users overwhelmingly perceive that some of the most popular LLMs have left-leaning political slants.
>
> For 18 of the 30 questi
[1] https://www.gsb.stanford.edu/insights/popular-ai-models-show-partisan-bias-when-asked-talk-politics
Re: (Score:2)
Bearing in mind that most of the world thinks even left leaning democrats are right wing extremists. Maybe most chat-bots are also right wing in their view
I have decided to quit voting (Score:1)
Because no matter if the politician is democrat or republican they all have dirt on them, so I refuse to vote because I refuse to endorse criminals
Cynical of ads (Score:1)
I think it's because people recognize ads as "trying to sell you on something/some idea" so they have their guard up more than they do for AI chatbots.
Re: (Score:2)
Nice on India/Russia/China/(Eastern-EU region)/Africa.
While your country still sucks, you take money to mess with others.
Re: (Score:2)
Lies? And 47 cancelled jobs and inflation and another major report, but he says everything's wonderful. And you believe him.
Sucker. I mean, unless you're a chatbot.