Note: We’ve transitioned from cimtech.solutions to cimtech.org for a unified mission-focused domain.
TL;DR
- What happened: Anthropic refused a government request to loosen AI safeguards.
- What followed: The Pentagon canceled the contract and labeled Anthropic a supply chain risk.
- Why it matters: This raises serious questions about AI ethics, privacy, and human oversight.
- Faith lens: Human dignity and meaningful human control must remain central.
- Latest update: Anthropic has sued the Pentagon over the designation.
Even if you don’t closely follow AI news, you’ve likely seen headlines about the recent dispute between Anthropic (the company behind the Claude AI chatbot) and the U.S. government, particularly the Department of Defense (DoD). The story has appeared across major news outlets worldwide.
To make sure we’re all on the same page, here’s a plain language summary based on my understanding of the situation.
What Happened?
Anthropic is widely regarded as one of the most safety focused AI companies in the industry and has historically ranked very high on AI safety indexes. According to public reporting, the Pentagon asked Anthropic to remove certain safeguards from Claude so it could be used for all “lawful” purposes.
Anthropic’s CEO, Dario Amodei, declined. He stated that Claude would not be used for mass domestic surveillance or fully autonomous weapons. Importantly, Anthropic has said these limitations were already part of the contract the Pentagon originally agreed to.
After this refusal, the government canceled Anthropic’s contract. President Trump then designated Anthropic as a supply chain risk, a label that has never been applied to a U.S. based company. This designation effectively prevents government contractors from using Claude in government work.
From the outside, the response feels disproportionate, almost like a playground argument that escalated far beyond what the situation warranted.
Industry Fallout
Shortly after Anthropic’s contract was canceled, OpenAI, the company behind ChatGPT, signed a contract with the Pentagon.
This move didn’t sit well with many observers. OpenAI had publicly expressed support for Anthropic’s ethical stance, yet then accepted the very deal Anthropic had declined. In response:
Claude rose to the #1 download spot in the Apple App Store
Over 4 million people reportedly signed up to boycott ChatGPT through the website quitgpt.org
Why Anthropic Says “No”
From a technical standpoint, it’s not even clear that Claude could be used in the way the government may want. Claude is trained using what Anthropic calls “Claude’s Constitution.”
Anthropic describes this constitution as a written set of values that guide how Claude behaves and makes decisions. In simple terms, it’s a way of embedding guardrails into the model from the ground up.
While I think Anthropic sometimes leans too far into anthropomorphizing Claude, I do appreciate the attempt to anchor AI systems in clearly stated values.
Two sections of the constitution seem especially relevant here:
Broadly safe — Claude should not undermine appropriate human oversight of AI systems, especially at this stage of development.
Broadly ethical — Claude should demonstrate honesty, avoid causing harm, and refrain from actions that are inappropriately dangerous.
Taken together, these principles likely make it impossible for Anthropic to simply “turn off” safeguards to meet the government’s request.
Ethics, Faith, and AI
This isn’t just a technical issue, it’s also a moral one, especially for those of us working at the intersection of faith and technology.
In 2019, the Ethics & Religious Liberty Commission (ERLC) published “Artificial Intelligence: An Evangelical Statement of Principles”, intended to help the church thoughtfully engage AI. Two sections are particularly relevant:
Article 8: Data & Privacy — While God knows all things, privacy remains an individual right. Governments and corporations should not violate it, and data collection should respect human dignity and the belief that we are made in God’s image (Imago Dei).
Article 10: War — The use of AI in warfare must be governed by love of neighbor. Humanity cannot delegate war entirely to machines, and meaningful human oversight is essential.
These principles align closely with what Anthropic appears to be trying to protect.
What Does This Mean for Us?
So where does this leave us as Christians living in the world, but not of it?
A few reminders:
AI is a tool, not a person. It is not intelligent in the way humans, created in God’s image, are.
Technology won’t save us. Scripture repeatedly warns against believing new tools will help us “make a name for ourselves” (see Genesis 11:4).
Banning AI outright doesn’t work. If we tell staff they can’t use AI at all, they’ll likely use it anyway, just without guidance or accountability.
A better approach is to provide:
A clear acceptable‑use policy
Approved tools
Guardrails that reflect both security best practices and Christian ethics
If you need help crafting an AI policy, feel free to reach out to CIM.
In Christ’s Name,
Jonathan Meester, VP & Chief Technologist, Computers in Ministry
Update: Anthropic has now sued the Pentagon, challenging the supply‑chain‑risk designation.