The Lone Wolf and the Death Star
Midjourney: Curt Doty
Dario Amodei Draws a Line.
There are moments when history turns not on policy—but on posture. This week, that posture belonged to Dario Amodei.
When reports surfaced that Anthropic’s Claude was allegedly deployed in a Defense Department operation tied to the capture of Nicolás Maduro, Amodei didn’t hedge. He didn’t issue a carefully lawyered shrug. He said the Pentagon violated the spirit—if not the letter—of Anthropic’s contract. And he reiterated what he’s long argued: AI should not be used for lethal operations or broad domestic surveillance.
That’s not just corporate positioning. That’s backbone.
In response, President Donald Trump, flanked by Defense Secretary Pete Hegseth, announced a ban on using Claude in any government program. The intent? Punishment. The effect? Promotion.
You don’t kneecap a product in Washington without making it a hero in Silicon Valley. Developers, startups, and civil society groups will now flock to Claude precisely because it was banned. In tech, exile is often branding.
Meanwhile, Sam Altman stepped into the vacuum. OpenAI signaled it has “red lines” similar to Anthropic’s—no autonomous weapons, no dystopian dragnet surveillance. Fine words. But words now collide with scale.
OpenAI has 900 million users. That’s not a startup. That’s a sovereign data state.
If ChatGPT can listen, understand, reason, and remember—who else can? If it becomes embedded in government contracts, what guardrails prevent user histories from being subpoenaed, pooled, or quietly analyzed? We’ve seen this movie before. The Patriot Act was sold as targeted security. It became Iraq Redux.
And let’s talk about the Orb.
Altman’s eyeball-scanning project—now rebranded as World, formerly Worldcoin—collects iris biometrics through a chrome sphere operated by Tools for Humanity. Iris data is permanent. You can’t reset your retina like a password. Critics warn that building a massive biometric database creates long-term risk if hacked, sold, or repurposed.
Biometrics are destiny. Once leaked, they’re forever.
If we’re worried about ChatGPT’s conversational memory, imagine adding irises to the mix.
And then there’s Ring.
Installed in roughly 33 million U.S. homes, Ring cameras can already be activated in coordination with law enforcement. The pitch is wholesome: find your lost dog Rover. The pretext is cute. The infrastructure is not. When millions of private cameras become a de facto surveillance grid, abuse isn’t hypothetical—it’s structural.
Which is why Amodei’s stance matters.
Yes, Anthropic recently softened its once hardline “automatic pause” safety pledge, shifting toward a more flexible roadmap. The gap between Anthropic and OpenAI has narrowed. Both now balance safety with competitiveness rather than treating safety as an absolute override.
But on weapons and mass surveillance, Amodei is still holding the line.
That deserves praise.
Director James Cameron warned us decades ago about autonomous systems slipping the leash. “The Terminator is an admonition,” he has said in various interviews, cautioning that we shouldn’t build what we can’t control. The metaphor isn’t subtle.
These are dangerous times. We cannot blindly trust government. We cannot blindly trust Big Tech. When one executive publicly resists pressure to weaponize intelligence at scale, we should notice—and support it.
Personally? I’m moving much of my workflow to Claude.
Not because it’s perfect. But because choices and voices are how we signal values in markets that increasingly resemble governments.
Big Tech pushback is risky.
It’s also an opportunity.
The future of AI won’t be decided by model size. It will be decided by spine. Thank you Dario.
Sources:
https://apple.news/AS-6waqVSQJSdn2f55J0_2A
https://www.hoschmorris.com/privacy-plus-news/worldcoin
https://www.wired.com/story/sam-altman-orb-eyeball-scan-launch-us/
https://www.cnn.com/2026/02/25/tech/anthropic-safety-policy-change
About the Author
Curt Doty is a former studio executive and award-winning creative director with deep leadership experience across the entertainment and branding industries. Ten years in Television. Ten Years in Movies.
As the founder of CurtDoty.co, a creative consultancy, Curt has led integrated marketing, multi-channel storytelling, branding, identity, and user experience initiatives for a diverse roster of clients.
Over the past 15 years, Curt has leaned into innovation—leading R&D projects at Apple, Toshiba, and Microsoft, and pioneering interactive content.
Today, Curt’s work also explores the intersection of AI and entertainment. A sought-after fractional leader (CCO, CMO), speaker, and AI educator, he focuses on demystifying AI for creatives and executives alike.
Curt is a sought after public speaker having been featured at Mobile Growth Association, Mobile World Congress, App Growth Summit, Promax, CES, CTIA, NAB, NATPE, MMA Global, New Mexico Angels, PRSA, EntrepeneursRx, Digital Hollywood, SHRM, Streaming Media NYC, and Davos Worldwide. Download his speaker presskit here.
Through public speaking, keynotes and podcasts, Curt is continuing his role as a visionary voice in the future of creativity. He is now a board member of The Human AI Innovation Commons, Encoding Equity Into AI-Generated Prosperity. A framework for ensuring the innovations arising from Human – AI collaborations benefit humanity broadly, not just corporate shareholders.
