OpenAI is introducing restrictions in response to global elections

OpenAI is taking measures in response to the elections that will take place in various countries in 2024. The research institute behind ChatGPT and DALL-E will no longer allow users to create chatbots that can imitate real people and institutions.

The institute reports that it also does not allow apps through its APIs that can be used for political campaigns or lobbying. Apps that could prevent or deter people from voting will now also be banned. The institute notes examples where voting procedures or requirements may be misrepresented.

As for fake images says OpenAI that it is working on tools that can detect the origin of images. This is already being experimented with within the company and the first tests are said to be ‘promising’. OpenAI will soon make the tool available to a group of testers, including journalists, internet platforms and researchers. They can then provide feedback.

ChatGPT users can now also report possible violations. They will also increasingly see results based on real-time global reporting. The language model will therefore always add source references and links to the answers.