This Week in AI: Microsoft's Says AI Can Diagnose Patients Better Than Doctors

One of the most promising uses for artificial intelligence comes from the world of medicine. Theproblem: there's a severe physician shortage, with the Association of American Medical Colleges estimating a deficit of up to 124,000 physicians by 2034, including 48,000 primary care doctors. The answer may come from AI.

Microsoft said it's taken "a genuine step toward medical superintelligence," according to a new report from Wired, which said the company's latest AI tools can diagnose patients "four times more accurately than human doctors." Interestingly, Microsoft's tools work directly with OpenAI's ChatGPT, Google's Gemini, Anthropic's Claude, Meta's Llama and xAI's Grok. Wired described this as similar to mimicking "several human experts working together."

Perhaps even more promising, Microsoft said its system can be configured to work within specific cost constraints, allowing for "explicit exploration of the cost-value trade-offs inherent in diagnostic decision making." The result, Microsoft said, is that its system often selected less-expensive tests and procedures other physicians would have, but still came away with "higher diagnostic efficiency."

This is just the latest time we've heard about the promise of AI in medicine. Previous studies have found that when AIs are paired with radiologists, they more accurately detect cancer than when either are working independently. And if there isn't a doctor readily available, such as in talk therapy, early signs indicate that AIs can help alleviate need there too, at least in the short term and under careful supervision.

"This is just a first step," Microsoft wrote. "We strongly believe that the future of healthcare will be shaped by augmenting human expertise and empathy with the power of machine intelligence."

Cloudflare has an idea for how content can make money from LLMs

Cloudflare, one of the biggest internet infrastructure companies, is launching a new initiative designed to help content makers block AI bots from crawling their sites, and instead charge for access. The idea is called "Pay Per Crawl," and TechCrunch reports it will allow website owners to charge a set rate. Alternatively, website owners can choose to block or allow AI crawlers all together.

"At scale, Cloudflare’s marketplace is a big idea that could offer publishers a potential business model for the AI era — and it also places Cloudflare at the center of it all," TechCrunch wrote.

Cloudflare's experiment comes at a critical time for musicians, artists, writers and other content producers, who are seeing traffic to their work decline as AI use across the web ticks up. The resulting loss of ad revenue has been significant for some publishers, contributing to sudden shutdowns of long-running publications, including Laptop Magazine this week.

While it's not certain how well Cloudflare's approach will work, it's clear publishers will need to change the way they do business on the web. Courts are increasingly saying that AI companies can use copyrighted works to train their technologies, so long as they're legally obtained. That means the only way content producers may be able to protect the value of their work would be to charge AI for access to their work at the outset.

ChatGPT referrals picking up

While companies struggle with the future of how AI may impact their businesses, they are increasingly finding that apps like ChatGPT are actually sending them traffic. A new report from Similarweb found that while "organic traffic" to news sites has dropped 26% since Google's AI Overviews launched last year, referrals from ChatGPT have skyrocketed more than 25x. Unfortunately, the tradeoff isn't enough to make up the difference, and content makers across the web are still reporting significant drops in traffic.

"Solutions to the news publishers’ crisis are few and far between," TechCrunch reports. OpenAI CEO Sam Altman recently acknowledged to the New York Times that jobs will likely be replaced by AI, leading to "real pain here, in many cases."

State AI laws to remain, for now

As lawmakers debated a spending bill over the past couple months, one theme that kept coming up was AI. Specifically, some lawmakers had attempted to include a provision in the large spending package that would bar states from creating individual rules around AI technologies for the next decade, effectively guaranteeing protections for tech companies who'd only have to worry about potential federal rules if the provision were to become law.

The move would also kneecap some state efforts around AI rules, including disclosure rules in advertisements, privacy protections for people's personal data, and others.

The provision was tossed before the Senate voted on the bill, however, so states are still free to set their own rules, for now.

More from MC News

• Keyboard 101: Intro to Computer Keyboards
• Fix It Yourself: Talking to iFixit on Why Repairable Tech Matters
• How to Use an AI Agent
• Scenes from the Micro Center Santa Clara Grand Opening
• How We Built THE ULTIMATE DOOM PC
• Chain of Thought: AI's New Reasoning Revolution
• Windows 10 End of Support: What You Need to Know

Ian Sherr is a widely published journalist who's covered nearly every major tech company from Apple to Netflix, Facebook, Google, Microsoft, and more for CBS News, The Wall Street Journal, Reuters, and CNET. Aside from writing, he tinkers with tech at home, is a longtime fencer -- the kind with swords -- and began woodworking during the pandemic.