A recent named the jobs most at risk of being replaced by artificial intelligence. The list includes journalists, news analysts, proofreaders and technical writers — all workers who verify and communicate accurate information to the public.
Tanu Mitra, an associate professor at the University of Washington Information School, calls these and other individuals “information integrity workers.” She recently was awarded a $649,971 grant from the National Science Foundation to study how large language models will affect these information workers.
The Faculty Early Career Development (CAREER) award, titled “Enabling Healthier Information Ecosystem: Understanding and Supporting Online Information Integrity Work in the LLM-Era,” will support Mitra’s work over five years, during which she will identify common challenges and needs from information workers who are most affected by the advances in generative AI, such as large language models (LLMs). The research, in particular, will focus on developing tools and frameworks to help frontline information professionals — journalists, fact-checkers, humanitarian responders, and content moderators — confront new threats posed by LLMs.
She chose a wide range of professionals to ensure the project has as large of an impact as possible.
“The technology has not caught up to help these information workers or information integrity workers to do their work,” she said, “And so the overarching goal of the proposal is to serve them and to empower people who protect the health of our information ecosystem.”
The CAREER award is one of NSF’s most prestigious and is given in support of early-career faculty who have the potential to serve as academic role models in research and education.
Mitra said the process of applying for the grant was long and stressful, involving extensive peer review and a long wait to learn whether the project was accepted.
“It’s fulfilling to know that the academic community and the research community thought that this work was important enough to get funded,” she said. ”LLMs are a really fast moving space in terms of research, and I'm really excited and proud that the topic would be relevant even if the technology has changed.”
As artificial intelligence tools increasingly threaten people’s jobs, Mitra and her team aim to design systems that integrate, rather than replace, human experience. By drawing on lived experience and expertise, they hope to address the shortcomings of current LLM models.
Once they can properly identify the challenges information integrity workers are facing, the researchers can build the specific tools that will help workers proactively respond to LLMs. Mitra wants them to have the ability to identify the vulnerabilities in these models, auditing them and finding the scenarios where they produce false information.
Mitra’s second goal is to integrate such tools into professionals’ day-to-day jobs, giving them agency over the information that is generated from the models.
“This will allow people who are doing information integrity work to have a more proactive approach to defending the information space, instead of having a more reactive approach,” she said.
One major part of the project’s implementation is education. Mitra will be working with , a Seattle-based education and research organization, to engage local schoolteachers in studying the risks and challenges of integrating generative AI in the classroom. Through this collaboration, Mitra’s research team can offer guidance along with practical tools to help both students and educators recognize information accuracy when using AI.
Mitra’s team will also collaborate with information workers worldwide through workshops and community outreach webinars. The project will extend through May 2030.
‘[This outreach will go] beyond the grant and beyond this five-year timeline,’ she said. ‘We can actually leave something that can have a broader and bigger impact.”
Pictured at top: Tanu Mitra speaks with students during the RAISE Winter Exposition earlier this year.