Medical researchers in South Dakota and across the United States hope to use artificial intelligence (AI) to treat cancer, predict the onset of Alzheimer's and diabetes earlier, diagnose and address disparities in the impact of kidney disease, and more.
But first we need data.
Large amounts of data.
Biotechnology Panel: Patient needs come first when implementing artificial intelligence and advanced technologies
Data held by government agencies, hospitals and other medical facilities, pharmaceutical companies, research institutions, and insurance companies.
Connecting researchers with data to train and improve AI medical applications must be done in a way that protects people's privacy and includes data from underserved and vulnerable populations, two leaders of the National Institutes of Health (NIH) told lawmakers this week.
Representatives from the NIH spoke online on Wednesday to members of a congressional study committee on artificial intelligence and regulating internet access for minors.
“Data is what drives artificial intelligence,” said NIH's Susan Gregulick. “We need real-time, high-quality data that's relevant to individuals and patients.”
According to Gregulick, the NIH has committed roughly $1 billion to research and development in machine learning and AI medical research since 2019, and plans to spend $296 million in 2023 alone.
University of South Dakota professor Bill Harris has received several NIH grants over his career, the most recent of which is a $506,000 grant to support research into an AI-powered model that looks at patterns of fatty acids in blood samples to predict the likelihood that a patient will develop Alzheimer's disease about four years sooner than doctors currently do.
Dr. Harris is a professor at the University of South Dakota School of Medicine, but also works in AI research through his company, OmegaQuant.
Another South Dakota project is part of the NIH Artificial Intelligence/Machine Learning Consortium to Advancing Health Equity and Researcher Diversity (AIM-AHEAD) program. In this project, South Dakota State University professor Semhar Michael is studying how machine learning can be used to identify health disparities among people with end-stage renal disease.
The SDSU project, which involves researchers from Dakota State University, Sanford Health and other researchers from within South Dakota and beyond, was awarded more than $1 million in grant funding over two years.
NIH's Gregulick said the SDSU project is one of 274 projects nationwide funded through the AIM-AHEAD initiative.
Gregulick said AI research in healthcare is still in its early stages. The NIH expects to see more “multimodal” AI projects in the future. These efforts will aim to integrate AI intelligence data generated by analyzing things like blood and tissue samples with other sources, such as speech recognition data. Speech recognition data could help triangulate the trajectory of cognitive decline and provide earlier interventions for Alzheimer's and dementia patients by training AI models to detect changes in speech patterns.
Privacy concerns
But without guardrails to protect patient data, no research can be conducted ethically, according to Lyric Jorgenson, NIH's associate director for science policy.
Lawmakers studying internet use by kids and AI face 'information overload'
It is important to anonymize data by removing names and other personally identifiable information, and to obtain patient consent and assistance at the point of data collection.
“We want to understand the risks of sharing information, especially with people who shouldn't have access to it,” Jorgenson said. “Think about putting it in a safe and making it accessible only to family members.”
She said communication with patients and the community will be key, and that the NIH is funding research into data use, as well as outreach and education activities about AI and data collection, but that basic data security practices will also need some tweaking.
Part of this will be storing data in the cloud, rather than on flash drives that can be passed from person to person.
Data in the cloud is segregated and made available only to vetted individuals, so administrators of cloud datasets can see who is accessing it and when, Jorgenson said. Data that's downloaded and stored on a hard drive or flash drive can't be tracked and managed with that level of precision, he said.
Republican Sen. Mike Rounds of South Dakota told state lawmakers that the U.S. is in a better position to protect privacy than countries like the United Arab Emirates, Saudi Arabia and China, which use surveillance systems to collect data to spy on its citizens.
“Is that where you want AI to be developed?” Lowndes asks.
Proposal for state legislators: Encourage AI development in South Dakota
Rounds has been a leader on AI issues in the Senate and is part of a bipartisan group of lawmakers who have met with tech industry leaders aiming to inform other lawmakers about the technology's benefits and potential pitfalls.
AI research, particularly in the area of data quality, represents an opportunity for young South Dakotans, he said, because an AI model is only as good as the data used to train it.
Schools such as Dakota State University, which has invested heavily in cybersecurity and other data science programs, are well positioned to do that work.
“These huge databases have to be accurate,” Lowndes said.
Q&A: The debate over regulating AI will impact every aspect of our lives
Funding medical research seems expensive, but so is managing many of the diseases that AI research could prevent or treat with the right investment, he said.
“It's not cheap, but it's a tiny investment compared to what you would pay to suppress and prevent the disease it's meant to treat,” Lowndes said.
The councillors in the research group wanted to know what they could do to move the research forward.
“You said we can't slow the progress of AI, but at the same time, we need policies, oversight and maybe training to govern its adoption,” said Rep. Chris Carr, a Republican from Sioux Falls.
Lowndes said the main goal should be to “stay ahead” in AI development. Supporting education programs focused on AI systems and maintaining a business-friendly atmosphere will help position the state to take advantage of new technologies, Lowndes said.
“I'm going to do everything I can to encourage the development of AI databases in all areas here,” Lowndes said, “so that's the catalyst for other things to happen.”
Lowndes said companies looking to develop data centers or AI hubs “will go to places that make it as easy as possible to do business.”
Get morning headlines delivered to your inbox
Subscribe
