https://www.theguardian.com/technology/2025/dec/04/chatbots-sway-political-opinions-substantially-inaccurate-study [p] Skip to main contentSkip to navigation Close dialogue1/1Next imagePrevious imageToggle caption Skip to navigation Print subscriptions Newsletters Sign in US[ ] * US edition * UK edition * Australia edition * Europe edition * International edition The Guardian - Back to homeThe Guardian [ ] * News * Opinion * Sport * Culture * Lifestyle Show moreHide expanded menu * [ ]News + View all News + US news + US politics + World news + Climate crisis + Middle East + Ukraine + US immigration + Soccer + Business + Environment + Tech + Science + Newsletters + Wellness * [ ]Opinion + View all Opinion + The Guardian view + Columnists + Letters + Opinion videos + Cartoons * [ ]Sport + View all Sport + Soccer + NFL + Tennis + MLB + MLS + NBA + WNBA + NHL + F1 + Golf * [ ]Culture + View all Culture + Film + Books + Music + Art & design + TV & radio + Stage + Classical + Games * [ ]Lifestyle + View all Lifestyle + The Filter + Wellness + Fashion + Food + Recipes + Love & sex + Home & garden + Health & fitness + Family + Travel + Money * Search input [ ] google-search Search + Support us + Print subscriptions + Newsletters + Download the app * + Search jobs + Digital Archive + Guardian Licensing + Live events + About Us + The Guardian app + Video + Podcasts + Pictures + Inside the Guardian + Guardian Weekly + Crosswords + Wordiply + Corrections + Tips * Search input [ ] google-search Search + Search jobs + Digital Archive + Guardian Licensing + Live events + About Us * UK * US politics * World * Climate crisis * Middle East * Ukraine * Football * Newsletters * Business * Environment * UK politics * Science * Tech * Global development * Obituaries Smartphone displaying a range of AI apps [ ] The study said tweaking a model after its initial phase of development was an importand factor in making it more persuasive. Photograph: Jono Searle/AAP View image in fullscreen The study said tweaking a model after its initial phase of development was an importand factor in making it more persuasive. Photograph: Jono Searle/AAP Artificial intelligence (AI) Chatbots can sway political opinions but are 'substantially' inaccurate, study finds 'Information-dense' AI responses are most persuasive but these tend to be less accurate, says security report Dan Milmo Global technology editor Thu 4 Dec 2025 14.00 EST Share Chatbots can sway people's political opinions but the most persuasive artificial intelligence models deliver "substantial" amounts of inaccurate information in the process, according to the UK government's AI security body. Researchers said the study was the largest and most systematic investigation of AI persuasiveness to date, involving nearly 80,000 British participants holding conversations with 19 different AI models. The AI Security Institute carried out the study amid fears that chatbots can be deployed for illegal activities including fraud and grooming. The topics included "public sector pay and strikes" and "cost of living crisis and inflation", with participants interacting with a model - the underlying technology behind AI tools such as chatbots - that had been prompted to persuade the users to take a certain stance on an issue. Advanced models behind ChatGPT and Elon Musk's Grok were among those used in the study, which was also authored by academics at the London School of Economics, Massachusetts Institute of Technology, the University of Oxford and Stanford University. Before and after the chat, users reported whether they agreed with a series of statements expressing a particular political opinion. The study, published in the journal Science on Thursday, found that "information-dense" AI responses were the most persuasive. Instructing the model to focus on using facts and evidence yielded the largest persuasion gains, the study said. However, the models that used the most facts and evidence tended to be less accurate than others. "These results suggest that optimising persuasiveness may come at some cost to truthfulness, a dynamic that could have malign consequences for public discourse and the information ecosystem," said the study. On average, the AI and human participant would exchange about seven messages each in an exchange lasting 10 minutes. It added that tweaking a model after its initial phase of development, in a practice known as post-training, was an important factor in making it more persuasive. The study made the models, which included freely available "open source" models such as Meta's Llama 3 and Qwen by the Chinese company Alibaba, more convincing by combining them with "reward models" that recommended the most persuasive outputs. Researchers added that an AI system's ability to churn out information could make it more manipulative than the most compelling human. "Insofar as information density is a key driver of persuasive success, this implies that AI could exceed the persuasiveness of even elite human persuaders, given their unique ability to generate large quantities of information almost instantaneously during conversation," said the report. Feeding models personal information about the users they were interacting with did not have as big an impact as post-training or increasing information density, said the study. Kobi Hackenburg, an AISI research scientist and one of the report's authors, said: "What we find is that prompting the models to just use more information was more effective than all of these psychologically more sophisticated persuasion techniques." However, the study added that there were some obvious barriers to AIs manipulating people's opinions, such as the amount of time a user may have to engage in a long conversation with a chatbot about politics. There are also theories suggesting there are hard psychological limits to human persuadability, researchers said. Hackenburg said it was important to consider whether a chatbot could have the same persuasive impact in the real world where there were "lots of competing demands for people's attention and people aren't maybe as incentivised to sit and engage in a 10-minute conversation with a chatbot or an AI system". Explore more on these topics * Artificial intelligence (AI) * Chatbots * Computing * Internet safety * news Share Reuse this content More on this story More on this story * [1024] Google's AI Nano Banana Pro accused of generating racialised 'white saviour' visuals 3h ago * [3750] The AI boom is heralding a new gold rush in the American west 10h ago * [1679] Anti-immigrant material among AI-generated content getting billions of views on TikTok 2d ago * [4054] Computer maker HP to cut up to 6,000 jobs by 2028 as it turns to AI 26 Nov 2025 * [5020] Leading law firm cuts London back-office staff as it embraces AI 21 Nov 2025 * [4104] Asos turns to AI stylists to win back shoppers after sales slide 12% 21 Nov 2025 * [3333] Crypto market sheds more than $1tn in six weeks amid fears of tech bubble 18 Nov 2025 * [5000] Elon Musk's Grok AI tells users he is fitter than LeBron James and smarter than Leonardo da Vinci 21 Nov 2025 * [2361] Don't blindly trust everything AI tools say, warns Alphabet boss 18 Nov 2025 Most viewed Most viewed 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. * UK * US politics * World * Climate crisis * Middle East * Ukraine * Football * Newsletters * Business * Environment * UK politics * Science * Tech * Global development * Obituaries * News * Opinion * Sport * Culture * Lifestyle Original reporting and incisive analysis, direct from the Guardian every morning Sign up for our email * About us * Help * Complaints & corrections * Contact us * Tip us off * SecureDrop * Privacy policy * Cookie policy * Tax strategy * Terms & conditions * All topics * All writers * Newsletters * Digital newspaper archive * Bluesky * Facebook * Instagram * LinkedIn * Threads * TikTok * YouTube * Advertise with us * Guardian Labs * Search jobs * Work for us * Accessibility settings * Back to top (c) 2025 Guardian News & Media Limited or its affiliated companies. All rights reserved. (dcr)