A global leader in treating complex illnesses says artificial intelligence can help specialists detect pancreatic cancer up to three years before it can be clinically diagnosed.
FDA Commissioner Dr. Marty Makary, a prominent surgeon and public health researcher, recently told "The Dan Bongino Show" this is tremendous news for the 60,000 to 70,000 pancreatic cancer patients diagnosed in the U.S. each year.
"Researchers at Mayo Clinic say they can identify changes on a CAT scan now using AI up to three years before you develop the cancer," he said, adding that "right now there's no screening for pancreatic cancer."
Pancreatic cancer is only the 10th most common cancer, but it is the third-leading cause of cancer death in the U.S. Survival improves if the disease is caught early, but most patients are diagnosed after it has already spread.
Only about 13.3% of patients survive five years after diagnosis.
Bob Maginnis, an author and national security expert, points out that AI has access to every scan, every medical note, and every study taken or produced in the last decade or more.
"It's able to ascertain, amongst those that eventually became positive for the cancer, what the symptoms were and what the images were and what the tissue said years prior," he relays.
There are more than 100 different types of cancer, each with different causes, behaviors and treatments. No matter what the cancer, he says catching it early is the key to recovery.
"That allows them to predict and therefore to begin to aggressively treat a carcinogen that's at the initial stages of onset," Maginnis says. "This is going to be a phenomenal improvement in predicting future carcinogens."
He believes AI will be instrumental in treating all forms of cancer and other medical issues across the board.
"Guardrails" still needed
The Trump administration has made a 180 from the hands-off approach it was taking to AI and now wants to review and manage any artificial intelligence created by U.S. tech firms.
Considering how integral the technology is to national defense and the race with China, the White House is reportedly considering an executive order that would bring AI development under government control.
"They hit a brick wall, and they came to recognize that there are some genuine issues that need to be addressed," Maginnis says.
Last month, Anthropic introduced a new app called Mythos that exposed thousands of vulnerabilities in the software used by U.S. banks and government agencies.
On Tuesday, Anthropic CEO Dario Amodei said the U.S. has six to 12 months to fix the problems before China catches up.
The Center for AI Standards and Innovation, an office within the Commerce Department's National Institute of Standards and Technology, announced Tuesday that it will conduct pre-deployment evaluations and "targeted research" on AI models developed by the companies.
"This stuff is moving so rapidly," Maginnis notes. "In spite of our competition with the Chinese, we need to take a step back. Guardrails are appropriate."
Another reason the government wants AI oversight is the military applications the technology will control, affecting everything from target location and acquisition to weapons control to logistics.
He says the prospect of such a tool shows that artificial intelligence can be a blessing and a curse.
