OpenAI has a 'cartoon' solution to tackle AI deepfakes in US midterm election 2026

​​OpenAI is turning some election-related AI images into cartoons. Ahead of the US midterm elections, which is set to take place on November 3, the AI company is applying image stylisation to requests depicting election interference or scenes that could discourage voting. The safeguard replaces realistic-looking images with illustrations, reducing the risk that an invented scene is mistaken for evidence of something that happened.In its latest update, OpenAI described the outputs as “stylised – also known as cartoonified – illustrations, rather than photorealistic images”. The approach allows political illustrations and social commentary while making potentially misleading scenes less convincing as documentary evidence. OpenAI said its rules against election interference and discouraging voting still apply, including when the imagery is cartoonified.How does the cartoon safeguard workThe safeguard changes how certain election-related scenes are presented, including when users explicitly request a photographic style.In tests reported by CNET, ChatGPT generated cartoon images when asked to show Immigration and Customs Enforcement officers outside a polling location. Requests to make the scene photorealistic still produced illustrations. In another test, CNET could not get ChatGPT to produce a photojournalist-style image of a voting machine recording the wrong choice.Such fabricated scenes could be presented as evidence of intimidation or problems with voting equipment. The safeguard aims to make that deception harder without blocking election-related illustration altogether.“The problem with deepfakes isn’t social commentary; it’s deception. So our aim is to degrade the evidentiary value of an image, while preserving room for illustrations and social commentary,” Josh Lawson, OpenAI’s head of elections, told CNET. The goal is to “make our platform less useful to bad actors by adding another layer of friction,” he added.What other election measures has OpenAI introducedOpenAI’s October update also outlined changes intended to help ChatGPT users find reliable voting information. Its integration with Democracy Works is live in the US across languages, providing registration requirements, deadlines, voting locations, and other election details. Responses include links to official sources so users can check the information.After polls close in the US and Brazil, ChatGPT now provides live vote counts thanks to an integration with the Associated Press. OpenAI said election officials are the authority on voting rules and certified results.Through its Daybreak program, the company has also offered eligible voting-system manufacturers and state election authorities free access to models for cybersecurity work, including identifying weaknesses in election infrastructure.Does this change OpenAI’s political content rulesOpenAI continues to prohibit political advertising on its platform. It also prohibits using its products to create or distribute campaign messaging at scale that advocates for or against candidates, political parties, or ballot measures, including images and videos. Campaigns can still use its tools for human-directed work such as internal briefings, planning, translation, compliance and administrative tasks.The cartoon safeguard restricts realistic imagery in specified election-related situations. It does not make otherwise prohibited attempts to interfere with voting permissible.​You use AI every day. Now get your AI Quotient. Take the AIQ test.