ChatGPT to bring real-time vote counting in the 2026 Elections in Brazil

A OpenAI, famosa pelo ChatGPT, está se preparando para um papel inédito nas eleições de 2026 no Brasil e nos Estados Unidos. A empresa anunciou um pacote de medidas que promete combater a desinformação com o uso de inteligência artificial, e o destaque é a apuração de votos em tempo real. Durante as noites de votação, os resultados serão exibidos diretamente na interface do ChatGPT. A base para essa operação será o monitoramento oficial da Associated Press (AP), garantindo que as informações sejam precisas e atualizadas.

This initiative by OpenAI is no accident. There is growing pressure from global authorities to prevent generative technologies from being used improperly, unfairly influencing political debates. In Brazil, the use of synthetic tools in campaigns is already a reality, and a recent survey by the Election Observatory identified 18 artificial avatars acting as political influencers and commentators. This shows how much the digital environment can be manipulated, and OpenAI wants to be at the forefront of mitigating these risks.

In addition to vote counting, ChatGPT will also prioritize practical information on deadlines, polling locations, and registration procedures. In the United States, this database will be provided by Democracy Works. In Brazil, OpenAI is still searching for a definitive national partner, but cooperation with the Superior Electoral Court (TSE) is a possibility, given the track record of compliance by Big Techs in the country.

Another important point is the fight against deepfakes. OpenAI will apply the SynthID digital watermark to images generated by ChatGPT. Developed in partnership with Google, this mechanism helps prove the provenance and modification history of a digital file, even after compressions or edits. However, this technology only works for content produced by OpenAI, meaning it does not detect alterations from competing platforms like Meta AI or Midjourney.

And there's more: OpenAI is banning automated electoral advertising on its platform. Large-scale political marketing campaigns simulating organic engagement are out of the question. Paid political ads are also banned during the election cycle. The legitimate use of generative AI by campaign committees will only be allowed for internal tasks, such as translations and scheduling, always under human supervision.

These measures show that Big Techs are taking election security seriously. And with the growing use of virtual assistants in daily life, it is essential to choose the most efficient model for each demand. Detailed analyses can help you decide between ChatGPT and Gemini for your daily tasks.