OpenAI to Prevent People from Using AI for 2024 Elections Disinformation

OpenAI is making big changes to its AI use policy in preparation for the upcoming 2024 Elections.

The creator of ChatGPT and DALL-E announced in a blog post this Monday that it will prevent people from using its AI technologies to spread disinformation and promote transparency.

OpenAI said it will no longer allow its users to create chatbots and applications impersonating government bodies or dissuading people from attending election polls.

DALL-E will also have "guardrails" to prevent people from requesting generated images of real people, including politicians and candidates.

The AI firm said it is already experimenting with its "provenance classifier" to detect AI-generated images made using DALL-E.

Changes to DALL-E's uses will be part of OpenAI's digital verification effort Coalition for Content Provenance and Authenticity set to be implemented early this year.

Journalists, verified platforms, and researchers will be the first to test the new policies before it is released to the public.

AI Labeled as Threat Amid Growing Disinformation Online

With more than 40% of the world's population preparing to enter the election, AI-fueled disinformation is expected to ramp up in the coming months.

Political figures have become central figures in deepfakes and AI-generated images meant to manipulate people's perceptions.

One of the many deepfakes released just last year are the generated pictures of former President Donald Trump being forcefully arrested by police forces.

DALL-E has since blocked similar requests since late 2023 as something that "may not follow our content policy."

Top global leaders have already raised the issue brought by the emerging technology to democracy and policy making during the annual World Economic Forum in Davos, Switzerland.

For OpenAI's part, the tech upstart vowed to incorporate actual news articles and government election website links to answer user inquiries related to elections.

The AI firm has earlier signed a deal with Axel Springer SE, and later with Thomson Reuters, to allow access to its licensed contents and generate summaries via ChatGPT.

OpenAI Vows to Work with Credible Organizations

The announcement follows CEO Sam Altman's promise to make efforts to combat generative AI's ability to compromise the election process.

Altman testified in front of the US Congress last year to push for AI regulation amid its growing application and increasing pace of development worldwide.

OpenAI is already working with the National Association of Secretaries of State, one of the oldest nonpartisan watchdogs for politicians and public officials.

The AI firm has yet to announce if it will also work with other political watchdogs for other countries holding elections soon.