Last week, OpenAI unveiled its GPT Store for third-party creators of custom chatbots (GPTs) to list and ultimately monetize their creations. The company is not stopping there as it continues to make headlines in the first month of 2024. OpenAI recently published a blog post outlining its plans to implement new safeguards around its AI tools, particularly the DALL-E image generation model and ChatGPT’s information citations, in an effort to combat disinformation ahead of the upcoming wave of elections worldwide.
Protecting the Integrity of Elections
The blog post begins with a statement emphasizing the need for collaboration and the protection of the democratic process: “Protecting the integrity of elections requires collaboration from every corner of the democratic process, and we want to make sure our technology is not used in a way that could undermine this process.” OpenAI highlights several existing safeguards currently in place around its AI tools. Among these safeguards is a “report” function that allows users to flag potential violations of custom GPTs and their behavior. OpenAI states that impersonation of real individuals or institutions is against their usage policies.
OpenAI’s blog post also mentions that users of ChatGPT will gain access to real-time news reporting globally, including attribution and links. This aligns with the company’s previous partnerships with renowned news outlets such as the Associated Press and Axel Springer.
Implementation of Safeguards
One of the most noteworthy commitments is OpenAI’s partnership with the Coalition for Content Provenance and Authenticity (C2PA), a non-profit initiative by tech and AI companies to label AI-generated content and imagery with cryptographic digital watermarking. OpenAI plans to implement these credentials and digital watermarking on its DALL-E 3 imagery early this year. These measures will enable reliable detection of AI-generated content in the future. Additionally, OpenAI mentions the introduction of their “provenance classifier,” a new tool designed to detect images generated by DALL-E. This tool, which has shown promising initial results, will soon be available to a select group of testers including journalists, platforms, and researchers for feedback.
As political activists and organizations, like the Republican National Committee (RNC) in the U.S., are already utilizing AI to craft deceptive messaging and even impersonate rival candidates, the question arises: Will OpenAI’s tools be sufficient in countering the anticipated surge of digital disinformation? It remains difficult to determine, but OpenAI clearly aims to assert its commitment to upholding truth and accuracy, despite the potential misuse of its tools for malicious purposes.
