AI models can meaningfully sway voters on candidates and issues, including by using misinformation, and they are also evading detection in public surveys according to three new studies.
Scientists are raising alarms about the potential influence of artificial intelligence on elections, according to a spate of new studies that warn AI can rig polls and manipulate public opinion.
In a study published in Nature on Thursday, scientists report that AI chatbots can meaningfully sway people toward a particular candidate—providing better results than video or television ads. Moreover, chatbots optimized for political persuasion “may increasingly deploy misleading or false information,” according to a separate study published on Thursday in Science.
“The general public has lots of concern around AI and election interference, but among political scientists there’s a sense that it’s really hard to change peoples’ opinions, ” said David Rand, a professor of information science, marketing, and psychology at Cornell University and an author of both studies. “We wanted to see how much of a risk it really is.”
We need to understand here that the real danger isn’t that someone is going to put out a video of Trump saving an armload of orphans from a burning building and it will be so convincing that everyone believes it, or that someone is going to make a “secret recording” of Zohran Mamdani talking about plans to do nine-hundred and eleven more 9/11’s, the real problem in the short term is that this technology is going to inject a seed of doubt or plausible skepticism into everything we see, read and hear.
Do you work hard all day, all week and only have time for watching political coverage a half hour a week like so many Americans right now? If you don’t know shit about the people running against each other, and there are very convincing videos of each of them saying radical, crazy bullshit or beating up puppies, you’re not going to say “Hmn, this can’t all be real, I should dig deeper.” Nobody fucking does that. We have to stop pretending that the average person has any desire to sift through raw sewage for kernels of edible corn.
No, they will turn it all off. They will shrug and just vote randomly when the time comes.
This is very evident by our previous federal election cycle, where we had some of the highest voter and youth voter turnout in US history, yet exit polling showed over and over that most people had no idea what the candidate’s actual positions, history or qualifications were, most voters said they couldn’t decide between Harris and Trump. People who stood behind Sanders voted for Trump. People said that at least Trump seemed “genuine” which is really true. The man, despite lying and cheating and being a giant asshole, he does seem genuine about it, and people don’t know what else to turn to.
This is how AI and technology like algorithms and atomized perspectives are going to subvert democracy in the short term.
In the long term, advanced AI will be able to manipulate your thoughts and feelings like you’re a fucking chicken being hypnotized by drawing a line in front it’s beak. It will know all your backdoors, hacks and overrides for you specifically and be able to predict all your patterns and thoughts that you think are soooo unique and personal. We’re so fucked.
BREAKING: SCIENTISTS WORRIED THAT TOOL DESIGNED TO DO THE THING IT WAS DESIGNED TO DO, WILL IN FACT DO THAT.
Why do you think the tech bros are so erect over AI? One more tool to keep us under the whip and yoke of fascism.
Wait, you mean to tell me that the billionaire machine, built by billionaires to say things billionaires like, might influence politics in favor of billionaires? Say it ain’t so
*again
I’m fairly certain that election fraud is already rampantly being practiced by various factions. I could post a long diatribe of sources for that, but the idea that AI manipulation of social media is what will push us over the line seems quaint.
I read that as “Scientists are increasingly worried that people are generally stupid and cannot confirm or distinguish truth from lies”. And, dammit, they’re right!
Data aggregation and mining have already made a mockery of the idea of humans having free will. Why cry now?
We are just a herd. Can I move THAT sheep? Maybe. Can I move the herd? Absolutely.
Why not just have AI do the voting?
Billionaires who own the government and direct it to ignore the will of the people with Citizens United and gerrymandering.
Narrator: little did they realize, It already had.
The last federal election cycle was a constant deluge of very bad renderings of Trump saving babies from fires and all the old ladies and Joe Rogan listeners had no idea what was real or not, because the prior scourge, the attention-span-eroding algorithm has trained people to not look at anything for more than a half second.
These things aren’t even necessarily designed or used to make people think that Trump is actually a beefcake riding a tank, it’s supposed to make everyone doubt what’s real or not. That’s all they have to do, is get people to tune out and doubt everything they read and hear as “AI.” That’s the real power of this shit, it makes us not want to be involved because it forces people to triple check everything we see and hear.
And this is exactly what exit polling in 2024 implied.
Instead of massive advertising campaigns that promote the messages we’re “supposed” to hear from large corporations an financial backers?
How is AI not a couple large corporations with the world’s biggest assholes’ financial backing.
From the horses mouth
AI is already influencing elections — and will continue to — in several important ways. Here’s a clear, non-technical breakdown:
1. Voter Persuasion & Targeted Messaging
AI makes it easier to analyze huge amounts of data about voters and generate personalized political messages. Campaigns can use AI to:
- Predict which voters are persuadable
- Tailor messages to specific groups
- Automate outreach (emails, texts, ads)
This can increase efficiency but also raises concerns about manipulation and privacy.
2. Deepfakes & Misinformation
AI tools can generate:
- Fake videos of candidates
- Synthetic audio imitating their voices
- Hyper-realistic fake news articles or social posts
These can spread rapidly and confuse voters, especially close to election dates.
3. Content Moderation & Detection Tools
AI is also part of the solution:
- Platforms use AI to detect bots, deepfakes, and coordinated misinformation campaigns.
- Fact-checking organizations use AI to analyze claims faster.
4. Automated Bots & Influence Campaigns
AI-powered bots can simulate real users and amplify political narratives at enormous scales. This can distort what appears to be “public opinion.”
5. Voter Information & Assistance
Positive uses include:
- Chatbots that provide voting information
- Translation tools increasing accessibility
- Analysis for journalists and researchers
Bottom Line
Yes — AI will influence elections, both positively and negatively. It can improve access to information and campaign efficiency, but also increases risks like deepfakes and targeted manipulation. Most experts argue the impact will grow, making transparency, regulation, and digital literacy increasingly important.
If you want, I can also explain:
- How deepfakes are detected
- What governments and platforms are doing to prevent election interference
- Ethical guidelines for using AI in political campaigns
Which LLM is this from, ChatGPT?
At one time I remember thinking, in my naïveté, that Americans actually valued scientists and progress. Silly me.
