The FY26 NDAA includes five AI provisions worth paying attention to — from logistics integration to AGI preparedness. We break them down:

AI has potential to reshape our armed forces, but history shows that premature, revolutionary force redesigns often fail. ARI's Morgan Plummer makes the case for measured adoption over hype in his latest War on the Rocks op-ed 👇 warontherocks.com/2025/12/the-ca…

The U.S. military has a well-earned reputation as a late adopter, but occasionally it leans forward & attempts to re-organize around new transformational tech. Despite conventional wisdom of “more AI, all the time” at DOW my latest in War on the Rocks explores why an AI-based force

An important PSA this holiday season

"Right now we are in a world where 4 or 5 people could crash the global economy." Nvidia alone is carrying the stock market. Data center buildouts are half a trillion this year. Brad Carson on the concentration risk behind the AI boom.

Me Dec 26th - Jan 1st:

POLL from Morning Consult: 74% of Michigan voters would be more likely to vote for candidates who prioritize protecting children from harmful AI content. 63% say they'd back a Senate candidate who supports stronger AI safeguards overall.

Wishing you a happy New Year from the team at ARI! 🎆

Why exactly should/must they accept that? Military sure is a context in which you want your systems to be reliable and aligned.

The tactics that won the Battle of Midway weren't improvised, they were refined through years of wargaming. Today, AI is compressing warfare timelines to milliseconds. US wargaming needs an update. Our latest Policy Byte analyzes the path forward ↓ ari.us/policy-bytes/m…

“We must accept that the risks of not moving fast enough outweigh the risks of imperfect alignment” Exactly what you’re looking for in high risk domains. 🧐🙃😒

"The fundamental disagreement about AI isn’t one about technology, but about values, said Daniel Cochrane. “It’s a disagreement over what it means to flourish as a human being and as a society of humans." This is the case.

AI is reshaping the cyber domain. This week we hosted Helen Toner, David Feith, and Lt. Gen. (Ret.) Jack Shanahan for a discussion on AI-enabled cyber and US-China strategic stability. For info on future events, subscribe for updates from ARI → ari.us/subscribe/

Calling any defense tech scientists (Anduril Industries et al)., for help on a website on AI weapons I've tried to capture the SOTA on AI usage in drones. Love to have you look at it and give me feedback. ai-weapons-site.vercel.app/technical.html

JUST ANNOUNCED: The AI Infrastructure Buildout: Bubble or Boom? Trillions are flowing into AI data centers, chips, and energy. Is this a productivity boom or a bubble? Feb 17 | 3 PM ET with economists from Goldman Sachs, Moody's, and Yale University. Register: zoom.us/webinar/regist…

Now this is an important story. It will be interesting to see how OpenAI and Google stand on this.

I propose: NIST should be tasked with developing a standardized protocol for AI agent identification and communication. In so doing, NIST should convene industry, civil society, and even seek international input.