ChatGPT creator launches plan against electoral misinformation in Brazil

OpenAI, creator of the artificial intelligence platform ChatGPT, launched this Wednesday (27) a plan against electoral disinformation in Brazil and the United States. The text explicitly talks about "protecting elections" against the spread of fake news in system-generated content.

Planning envisages an agreement with the Associated Press to provide real-time vote counting of elections in both countries. Additionally, the company aims to monitor ChatGPT's political neutrality and combat the malicious use of AI for political purposes.

One of the topics most discussed in the statement is deepfakes, a technology used to produce realistic videos in which people appear doing and saying things they have never done or said.

Continue lendo

28/05/2026

Chinese company says it has a device capable of "translating" pet sounds

Chinese technology company Hangzhou Meng Xiaoyi Technology claims to have developed a device capable of "translating"...

27/05/2026

Anahí and Poncho: project reveals reason for reunion between the two

Singer Anahí, 43, and actor Alfonso Herrera, 42, caught the attention of fans when they reunited in January after year...

27/05/2026

With Bernard decisive, Atlético-MG qualifies in the Copa Sudamericana

Atlético-MG is in the round of 16 of the Copa Sudamericana. This Wednesday (27), at Arena MRV, Bernard scored the goal...

"To help combat misleading 'deepfakes', we are investing in a multi-layered approach to digital provenance that will allow people to verify that the content they are seeing was created or modified with AI," the statement reads.

The idea is for users to have access to technological tools and check whether the images were generated by ChatGPT, Codex or the OpenAI API through digital watermarks. OpenAI, however, did not specify how these watermarks would be visible, whether they would only be applied to political figures, how this use would be identified and whether there would be a distinction between users of the platform's simple plan and premium plans.

Top Cifras

Toque agora.

1 Eu Não Tô Dizendo Tchau Wesley Safadão

2 Un Pato Natalia Lafourcade

3 Somente Nela Paulinho Moska

4 Luv Sick Part 3 Nujabes

The company also said it will work with social media companies to prevent the spread of false content, but did not explain with which or how. "As most people encounter AI-generated content on social networks and messaging apps, we are pleased to collaborate with social media companies as they evaluate measures to protect elections, using provenance markers as a relevant signal to decide which civic content to recommend and distribute in users' feeds."

Still in May, OpenAI closed a billion-dollar deal with the United States Department of Defense, hours after President Donald Trump banned Anthropic, its competitor that was placed on the government's blacklist after demanding security measures from the Pentagon for the use of its tool in wars.

Combating misuse
OpenAI also promised that it will work to prohibit the use of products "to create or distribute large-scale campaign messages for or against candidates, political parties, or ballot measures."

"We believe, however, that political campaigns should be able to use our tools for responsible human-driven work, such as internal briefings, planning, day-to-day writing, translation, compliance, and administrative tasks. These uses are permitted by our policies," the note says.

And he adds: "We believe it is important for people to be able to use AI systems to learn, explore and discuss political issues, and we will continue to allow this while preventing malicious uses by bad actors."

For the company, the elections are "an important moment for cyber defense professionals in various sectors". For them, AI has a "critical role in strengthening digital infrastructure", including in systems that support elections.

AI in elections: Brazil has 18 political influencers generated by artificial intelligence

Source: CNN