The National Institute of Standards and Technology (NIST), the U.S. Department of Commerce agency that develops and tests technology for the U.S. government, corporations, and the general public, announced Monday the launch of NIST GenAI, a new program led by NIST to evaluate productive AI technologies, including artificial intelligence that generates text and image.
NIST GenAI will release benchmarks, help build “content authenticity” (i.e. deepfake-checking) detection systems, and encourage software development to identify the source of fake or misleading AI-generated information, NIST explains in the recently launched NIST GenAI website and in a Press release.
“The NIST GenAI program will issue a series of challenge problems [intended] to assess and measure the capabilities and limitations of artificial intelligence production technologies,” the press release states. “These assessments will be used to identify strategies to promote information integrity and guide the safe and responsible use of digital content.”
NIST’s first GenAI project is a pilot study to create systems that can reliably tell the difference between human-generated and AI-generated media, starting with text. (While many agencies aim to detect deepfakes, studies and our own testing have shown that they are patchy at best, particularly when it comes to text.) NIST GenAI invites teams from academia, industry, and research labs to submit either “generators” — Artificial intelligence systems for content generation — or “discriminators”, which are systems designed to recognize content generated by artificial intelligence.
Authors in the study must create abstracts of 250 words or less, providing a topic and a set of documents, while adjudicators must identify whether a given abstract is potentially AI-written. To ensure fairness, NIST GenAI will provide the necessary data to test the generators. Systems trained on publicly available data and those not[comply] with applicable laws and regulations” will not be accepted,” NIST says.
Registration for the pilot will begin on May 1st, with the first round of two scheduled to close on August 2nd. Final results from the study are expected to be published in February 2025.
The launch of NIST GenAI and the deepfake-focused study comes as the volume of AI-generated misinformation and disinformation grows exponentially.
According to data from Clarity, a deepfake detection company, 900% more deepfakes have been created and published this year compared to the same period last year. It causes alarm, understandably so. ONE recently voting from YouGov found this 85% of Americans were worried about the deceptive deepfakes that are spreading on the internet.
The release of NIST GenAI is part of NIST’s response to President Joe Biden’s executive order on artificial intelligence, which established rules requiring greater transparency from AI companies about how their models work and established a series of new standards, including flagging content generated by artificial intelligence.
It’s also the first announcement from NIST about artificial intelligence since the appointment of Paul Christiano, a former OpenAI researcher, to the agency’s AI Safety Institute.
Christiano was a controversial choice for his “doomerist” views. he once was foreseen that “there is a 50% chance that AI development will end up [humanity’s destruction].” criticswhich reportedly includes scientists at NIST, fear that Cristiano may encourage the AI Security Institute to focus on “fantasy scenarios” rather than realistic, more immediate dangers from artificial intelligence.
NIST says NIST GenAI will inform the AI Security Institute’s work.