AI newswatch: American AI in the United Arab Emirates, healthcare AI investment, med schools embrace AI, more

Buzzworthy developments of the past few days: 

• Here’s a big thing to watch for in the wake of the President’s flashy Middle East tour: the AI “acceleration partnership” between the U.S. and the United Arab Emirates. A “close collaboration” can be a nebulous concept to keep an eye on, but this one includes the building of the largest AI infrastructure campus outside the U.S. The conspicuous brick-and-mortar compound is to be constructed by G42, an aggressive technology holding group based in Abu Dhabi that has a hand in numerous industries, including healthcare. The facility is to sprawl out over something like 10 square miles and consume copious gobs of energy from gas, nuclear and solar sources. As part of the partnership, the U.S. will allow the U.A.E. to buy half a million Nvidia chips a year, starting this year. The Commerce Department says the semi-constitutional monarchy has agreed to abide by U.S.-approved standards for technology oversight. Commerce adds that the AI acceleration campus will house U.S. hyperscalers and large enterprises “that can leverage the capacity for regional compute with the ability to serve the Global South.”
   
  • In the news site Arabian Gulf Business Insight, reporter Justin Harper considers the Trump Administration’s key motives for forging such a close and massive partnership on AI with a foreign government whose historic worldview is so unlike ours. “The long-term agreement positions Washington as a major player in shaping how AI is deployed across the Gulf, offering a strategic counterweight to growing Chinese and Russian tech influence in the region,” Harper writes. “It sends a clear message: The U.S. intends to drive innovation alongside trusted allies, while reinforcing its technological and geopolitical edge against rival powers.” Harper also quotes Secretary of Commerce Howard Lutnick, who bluntly called the agreement with the U.A.E. “a major milestone in achieving President Trump’s vision for U.S. AI dominance.” 
     
• Cathie Wood is bullish on technology coming out of the U.S. in this, the second Trump era. And when Cathie Wood talks investment, people do tend to listen. This week the influential CEO of Ark Invest told Bloomberg Television that no economic sector is more exciting a playing field for AI innovators than healthcare. “I think healthcare is the most underappreciated beneficiary of AI out there,” she said. “Robotics, energy storage, AI, blockchain technology and multiomic sequencing in the life science space are going to be the big drivers in the years ahead.” Summary coverage by MarketWatch here. 
   
• Two years ago you could have dumfounded a lot of doctors asking them to distinguish a large language model from a large iced latte. And if that doesn’t inspire you to mull how far healthcare has come with AI, and how fast, think about this. Stanford School of Medicine appointed its first director of AI in medical education less than three months ago. The person who holds that title, data scientist Jonathan Chen, MD, PhD, tells AAMC News the technology is coming on so quickly that courses designed today might creak with age before a crop of students even completes them. “We—as an institution, as a profession—are going to be left behind if we don’t plan for where [medical AI] is going,” Chen says. “Because it’s coming at us really fast.” The headline of the piece says it even more succinctly: “Medical schools move from worrying about AI to teaching it.” Read it and say [bleep]. 
   
• And then there are medical call centers ‘staffed’ by AI. Human turnover at those workplaces is painfully high. By contrast, AI voice bots never burn out, much less walk off the job mid-shift. A KFF Health News article picked up by the Los Angeles Times looks at the pros and cons of the AI solution to the perpetual worker shortages ever plaguing the field. “Startups are marketing AI products with lifelike voices to schedule or cancel medical visits, refill prescriptions and help triage patients,” the piece points out. “Zocdoc, the appointment-booking company, has introduced an automated assistant it says can schedule visits without human intervention 70% of the time.” On the other hand, vocal AI can only ever simulate human contact, not replicate it. “AI tools don’t make medical decisions,” Kaiser Permanente spokesperson Vincent Staupe says. “Our physicians and care teams are always at the center of decision-making with our patients and in all our care settings—including call centers.”
   
• Multilingual AI can be surprisingly sensitive. At Montefiore Medical Center in New York City, tech wizards have fine-tuned talking bots to feel patients’ pain—or at least sound like they do—even across language barriers. “We have taught our AI to be more empathetic,” Chief Digital and Information Officer Deepesh Chandra said in a podcast aired by Deloitte this week. “We are seeing how AI can do a much, much better job” now than earlier on, Chandra explains, “maintaining the right tone, the right demeanor—all aspects of communication—with the non-English [speaking] population.”
   
• The digitization of healthcare proceeds apace. There are few surer gauges of the transformation’s impact than the 2025 CPT code set released by the American Medical Association last September. With 420 overall updates, the latest batch represents nothing less than a “roadmap for the future of healthcare delivery,” remarks Angela Comfort, MBA, a revenue exec at Montefiore Medical Center, in ICD-10 Monitor. “CPT descriptors now explicitly require coding professionals to identify the level of automation and provider input, creating a demand for more nuanced documentation from clinicians,” she writes. “Coding professionals must be trained to recognize when AI is involved and ensure the record meets the payer requirements regarding human oversight and clinical validation.” 
   
• In developing regions of the world, AI can make the difference between a long life and an early death. This is clear in a new study presented at this week’s scientific meeting of the European Society of Cardiology Congress in Serbia. Researchers described their work trying out AI for interpreting electrocardiograms, or EKG graphs, in Kenya. The gold standard is echocardiography, which uses ultrasound to create images of the heart. However, few have access to “echo” technology in poorer parts of the world. For the study, the team trained an algorithm to flag signs of left ventricular systolic dysfunction (LVSD), which can lead to heart failure, on EKGs. The technique identified LVSD in almost 20% of 6,000 adults. And in a head-to-head matchup against echo in a 1,400-patient subset, the more accessible technology aided by AI had excellent sensitivity, good specificity and a negative predictive value of 99.1%. Dr. Bernard Samia, senior study author and president of the Kenya Cardiac Society, commented: “Our study shows the potential utility of AI-ECG algorithms as a relatively low cost and scalable tool for screening for heart disease, including heart failure in at-risk populations, in resource-limited societies.”