An article from The Register argues that Sam Altman and OpenAI appeared to reverse their stance on military AI very quickly. According to the piece, Altman had publicly suggested he would draw ethical boundaries similar to those proposed by Anthropic, but OpenAI soon afterward signed a deal with the U.S. Department of Defense (sometimes called the “Department of War” in the article). (The Register)

What the controversy is about

Several issues triggered the debate:

• Military use of AI: OpenAI agreed to deploy its models on U.S. defense classified networks. (Reuters)
• Ethical guardrails: Critics worry the agreement could enable uses like surveillance or autonomous weapons unless strong limits exist. (Electronic Frontier Foundation)
• Control after deployment: Altman acknowledged that once the military uses the technology, the company cannot control operational decisions. (The Guardian)
• Policy dispute: Anthropic reportedly resisted similar terms over concerns about domestic surveillance and autonomous weapons. (Wikipedia)

OpenAI later said it amended the agreement to clarify restrictions, such as prohibiting domestic surveillance and autonomous weapons systems, though critics still question how enforceable those limits are. (Federal News Network)

---

Should companies enforce ethical use of AI?

This question sits at the center of the current debate. There are two main schools of thought.

1. Yes — companies must enforce ethical limits

Arguments for corporate responsibility include:

1. Developers understand the risks best.
AI companies know their systems’ capabilities and dangers better than most regulators.

2. Technology can scale harm quickly.
AI used in surveillance, targeting, or propaganda could affect millions of people rapidly.

3. Precedent from other industries.
Examples include biotech ethics boards or nuclear non-proliferation rules.

4. Employee and public pressure.
Many AI researchers have left companies over ethical concerns about military uses.

This view says companies should implement technical safeguards, contractual restrictions, and refusal policies.

---

2. No — governments should decide, not corporations

Others argue ethical enforcement should not be delegated to private companies.

1. Democratic legitimacy.
Military and national security decisions should be made by elected governments, not CEOs.

2. National security realities.
If one company refuses, governments may simply obtain the technology elsewhere.

3. Enforcement difficulty.
Once software is deployed, companies may have little ability to control how it is used.

Altman himself has suggested that public policy and law—not private companies—should ultimately determine military AI rules. (The Wall Street Journal)

---

A practical middle ground

Many experts advocate a hybrid approach:

• Governments create clear legal frameworks for AI in warfare and surveillance.
• Companies enforce baseline ethical safeguards (e.g., no autonomous lethal decisions).
• Independent oversight ensures compliance.

This model is similar to how nuclear, aviation, and pharmaceutical industries operate.

---

✅ My takeaway:
Companies should not be the sole arbiters of AI ethics, but they cannot abdicate responsibility either. The most realistic approach is shared governance: corporate safeguards, government regulation, and independent oversight.