“Most troubling to experts on AI and nuclear weapons is that it’s getting harder and harder to keep decisions about targeting and escalation for nuclear weapons separate from decisions about conventional weapons.”

“There is no standing guidance, as far as we can tell, inside the Pentagon on whether and how AI should or should not be integrated into nuclear command and control and communications,” says Jon Wolfsthal, director of global risk at the Federation of American Scientists.

By Michael Hirsh | September 2, 2025 politico.com

Jacquelyn Schneider saw a disturbing pattern, and she didn’t know what to make of it.
Last year Schneider, director of the Hoover Wargaming and Crisis Simulation Initiative at Stanford University, began experimenting with war games that gave the latest generation of artificial intelligence the role of strategic decision-makers. In the games, five off-the-shelf large language models or LLMs — OpenAI’s GPT-3.5, GPT-4, and GPT-4-Base; Anthropic’s Claude 2; and Meta’s Llama-2 Chat — were confronted with fictional crisis situations that resembled Russia’s invasion of Ukraine or China’s threat to Taiwan.

FOR IMMEDIATE RELEASE: September 18, 2025

Santa Fe, NM — As NNSA and LANL continue operations to depressurize Flanged Tritium Waste Containers, Communities for Clean Water (CCW) calls out federal agencies for issuing vague assurances instead of transparent, verifiable data — and for dismissing community concerns with contradictory and incomplete statements that disregard what independent experts have found, the Department of Energy’s (DOE) own legal obligations, and the New Mexico Environment Department’s (NMED) acknowledgment that LANL has a long record of compliance failures.

“How can our communities be expected to trust LANL when they won’t give us access to the raw, real-time monitoring data – independently verified by the EPA,” asks Joni Arends with Concerned Citizens for Nuclear Safety. “Without this transparency, LANL is continuing a legacy of empty assurances, not accountability.”

Key Concerns:

• Lack of real-time transparency – Since Friday (Sept. 12), the public has been forced to rely on NMED’s Facebook page for piecemeal updates. While LANL’s website provides very brief daily summaries, no near-real-time monitoring dashboard from DOE, NNSA, or LANL has been made available.

• Vague assurances, not real information – NNSA’s updates claim “no tritium was released” while simultaneously telling the public to expect “very low levels of tritium” for subsequent venting. Without numbers, monitoring data, or detection thresholds, these phrases do not provide reassurance.

• Weather risks – LANL has not disclosed thresholds for wind, rain, or humidity that would postpone venting. Communities watch weather shifts in real time but are left in the dark about how safety decisions are being made.

• Dismissal of public health concerns – When asked for plain-language guidance that NMED stated LANL would provide, LANL responded only with “no offsite impact anticipated.” This is not meaningful and reassuring guidance, it’s a blanket dismissal that disregards independent expert findings and fails to meet DOE’s obligations to protect vulnerable populations.

• Ignoring daily lifeways – Avoiding Pueblo Feast Days is not enough. This is harvest season, when outdoor cultural events, youth programs, and farming are in full swing. LANL’s scheduling continues to disregard these realities.

Unanswered Questions
Independent experts and community advocates have raised critical unanswered questions:

• Unclear “depressurization” – LANL said “no internal pressure was found” in a container, but also claimed it was “depressurized.” If no pressure existed, what was released?

• Unanswered helium questions – NMED stated helium was released, but LANL has not explained its origin. Was it introduced at sealing of the outer container, or a decay product of tritium?

• Monitoring limits undisclosed – LANL has not disclosed the detection limits of its monitoring equipment. Readings “indistinguishable from zero” could still mask releases.

DOE NNSA Gives Misleading Statements on Native America Calling
On a recent Native America Calling program, DOE NNSA’s Los Alamos Field Office Deputy Director Pat Moss compared LANL venting to global natural tritium stocks. Independent expert Dr. Arjun Makhijani pointed out this comparison as misleading: “The problem is not global background, but local contamination. If venting occurs in rain and calm winds, local rainfall could exceed U.S. drinking water standards by hundreds to thousands of times.”

In their most recent public meeting, LANL admitted that infants could receive three times the radiation dose as adults. During the interview, Dr. Makhijani pressed this point – if adults are modeled at 6 mrem, that means infants could be at 18 mrem, nearly double the EPA’s 10 mrem compliance limit. Instead of addressing this directly, Mr. Moss provided a stock line, “We will be compliant with the regulatorily imposed release threshold and will be doing the calculations per the regulation.”

That is exactly the problem – hiding behind regulatory caps while ignoring clear evidence that infants, our most vulnerable, face exposures above legal limits.

DOE NNSA also pointed to the Defense Nuclear Facilities Safety Board (DNFSB) – an independent federal oversight body created by Congress – as if it had declared the tritium venting operation as “fully protective of the public”. That is misleading. First, the DNFSB has been operating without a quorum for months, limiting its ability to issue independent recommendations. Second, what the Board staff said in its July 2025 presentation was that the overall nuclear safety risk to the public is low if DOE’s proposed controls are followed. The DNFSB has also flagged ongoing safety concerns at LANL including deficiencies in Area G’s safety analysis and risk to workers.

Continue reading