In May 2025, I joined more than 160 blood cancer advocates on Capitol Hill. At the time, Congress was considering changes to Medicaid, a state and federal program that is a lifeline for more than 75 million Americans, including older adults, children, veterans, people with disabilities, pregnant people, and adults with low incomes. Together, we advocated against policies that threatened to put health care out of reach for millions.
One of those concerning policies was work reporting requirements for people covered by Medicaid. The then-proposed legislation mandated that states verify Medicaid enrollees’ monthly work status at least every six months to ensure that they are engaging each month in a minimum of 80 hours of community service, educational activities, and/or work. But undergoing cancer treatment is already a full-time job; many patients are unable to work because of their condition.
Patients are getting used to being recorded in the doctor’s office. More than a quarter of U.S. practices now use AI-based listening tools called ambient scribes, which capture visits in real time and draft clinical notes for clinicians to enter into patients’ medical records.
But what happens when it’s the patient doing the recording?
Nearly every patient now has a recording device in their pocket, and commercial large language models have made it easier than ever for them to transcribe and interpret audio files. A growing number of consumer apps — with names like VisitRecall, Advoca Health, and AlignCare — are using AI to do just that, delivering summaries and action items for patients after the fact. One called Kin Health, with backing from the co-founder of GoodRx, announced a $9 million seed round in May.
Interventions: Procedure: Endoscopic Enucleation of the Prostate; Procedure: Endoscopic Enucleation of the Prostate; Procedure: Endoscopic Enucleation of the Prostate
This is today’s edition of The Download, our weekday newsletter that provides a daily dose of what’s going on in the world of technology.
5 key points in Trump’s new AI order
Less than two weeks after scrapping an executive order on AI, President Donald Trump signed a new one on Tuesday. Promising to promote innovation and security, the policy represents a turning point in the White House’s AI governance—but is likely to attract criticism from both opponents and supporters of stricter regulation. Here are five key points from the order:
1. It’s created a voluntary review system: tech companies will be asked to share frontier models with the government for review 30 days before they plan to release them. 2. There’s no mandatory licensing: the government will not require permits before software can be deployed. 3. It establishes a dedicated AI cybersecurity clearinghouse: the new hub will coordinate security checks with the private sector. 4. It’s a watered-down version of the order Trump shelved last month: the earlier version requested models 90 days before their release. 5. But it’s still a move towards stronger AI oversight: the policy marks a clear departure from the White House’s previous hands-off approach.
Quay Barnett, who leads the effort at Anduril following a career in the Army’s Special Operations Command, aims to optimize “the human as a weapons system.” His vision is cyborg-inspired: drones and soldiers will see together, share information seamlessly, and make decisions as one.
—James O’Donnell
This is our latest story to be turned into an MIT Technology Review Narrated podcast, which we publish each week on Spotify and Apple Podcasts. Just navigate to MIT Technology Review Narrated on either platform, and follow us to get all our new content as it’s released.
The must-reads
I’ve combed the internet to find you today’s most fun/important/scary/fascinating stories about technology.
1 President Trump has signed an AI order that expands model oversight The long-awaited executive order aims to mitigate security threats. (NYT $) +It asks companies to submit models voluntarily for tests before release. (NPR) + It’s a slimmed-down version of the order Trump shelved in May. (WSJ $) + And marks a strategic shift in his AI strategy. (Reuters $) + A war over AI regulation is coming to the US. (MIT Technology Review)
2 SpaceX plans to raise $75 billion in IPO at $135 per share The company intends to sell 555.6 million shares. (Reuters $)+ The fixed price breaks from the traditional IPO process. (Bloomberg $) + Morningstar says the valuation should be nearly 50% lower. (BI)
3 Meta has scaled back plans to track workers’ clicks and keystrokes to train AI All staff can pause it for 30 minutes, with some fully exempt.(The Information $) + The changes follow a fierce backlash to the tracking plans. (Reuters $) + AI is supercharging surveillance. (MIT Technology Review)
4 Microsoft wants to ‘make users addicted’ to its new AI assistant According tointernal documents for the “Scout” tool. (404 Media) + Microsoft launched the assistant on Tuesday. (TechCrunch)
5 Mathematicians fear that AI threatens their field A new declaration raises concerns about AI’s trustworthiness. (Ars Technica) + It arrives a week after OpenAI said it solved a famous math problem. (WSJ $) + A startup wants to change how mathematicians do math. (MIT Technology Review)
6 Scientists have found a way to supercharge computer worms with AI The worm could target any known flaw in the world’s computers. (NYT $) + AI supercharging scams. (MIT Technology Review)
7 Google must let UK publishers opt out of AI search features Online publishers can choose not to appear in the AI Overviews. (BBC) + Google is now testing features for sites to exit AI search. (Reuters $)
8 America’s data center build-out is falling way behind schedule 60% of those planned for completion in 2027 aren’t yet under construction. (WSJ $) + Nobody wants a data center in their backyard. (MIT Technology Review)
9 EVs are getting cheaper worldwide—except in the US The US is short on supportive policies and affordable Chinese EVs. (Rest of World)
10 The European Parliament is ditching Google for… Quant The French search engine is the new default on in-house computers. (Politico) + The switch comes amid a broader push to wean the EU off US tech. (FT $)
Quote of the day
“SpaceX’s valuation could be richer than a plate of dauphinoise potatoes.”
—Dan Coatsworth, head of markets at AJ Bell, tells CNBC that SpaceX’s IPO price looks overloaded with expectations.
One More Thing
Marseille’s battle against the surveillance state
Heading toward Marseille’s central train station, Eda Nano points out what looks like a streetlamp on the Rue des Abeilles. But this sleek piece of urban furniture is not a lamp. It’s a video camera, with a 360-degree view of the narrow street.
Nano, a 39-year-old developer, wants to make Marseille residents more aware that they’re being watched. She’s part of a growing group of activists resisting the rise of policing cameras in their hometown.
Jakob Uszkoreit went from inventing the “T” in ChatGPT to putting the “AI” in Alnylam’s RNAi therapeutics.
After over a dozen years at Google, where he’d co-authored the seminal “Attention is all you need” paper that laid the foundations for artificial intelligence models like GPT, Uszkoreit in 2021 left a job at Google Brain to start Inceptive Nucleics. The company is building “AI foundation models of life” — models that will hopefully learn so much about how biology operates that they will be reusable across many different tasks without having to be trained specifically to do them.
On Wednesday, Alnylam and Inceptive announced a new, three-year strategic collaboration worth up to $2 billion, with $30 million up front in cash and equity in the startup. The additional payments will be predicated on the co-discovered drugs achieving preclinical, regulatory, and commercial sales milestones. In 2025, Alnylam’s total revenues were $3.7 billion.
A fourth major health insurer is suing HaloMD over its use of the No Surprises Act’s arbitration process, arguing that the middleman deceived arbitrators by sending them a “sham letter” and misleading price data.
Highmark Health, a Pennsylvania-based Blue Cross Blue Shield licensee with over 7 million members, claims in a complaint filed June 1 in U.S. District Court in Western Pennsylvania that HaloMD and one of its clients, a neuromonitoring provider called Bromedicon, submitted more than 450 ineligible disputes with the company and won more than $3.9 million. Like the three Blue Cross plans before it, Highmark wants those awards tossed and its money returned.
“Defendants’ scheme is not a good faith attempt to obtain fair reimbursement,” Highmark’s complaint says. “It is a deliberate effort to wrongfully extract inflated payments from Highmark.”
