Hello John and Doves,
In this letter I wrote about
the Pentagon's ultimatum on Anthropic - turn over Claude or
else. If came out in the news that Claude was used in
the raid that captured Maduro. Currently, Claude is
the only AI model operating on the U.S. military's 'fully
classified system'. And the DoD wants full control of
Anthropic's AI technology. Last July, the Pentagon
gave Anthropic a $200 million dollar contract to develop AI
for U.S. national security. Claude was the first AI
used by the DoD. Elon Musk's Grok is being used by the
US military in a 'classified setting'.
And I included info on a
recent war game exercise using Chat GPT-5.2, Claude Sonnet
4, and Gemini 3 Flash. Claude one 'hands down'.
Anthropic, Claude and DoD
Ultimatum:
Anthropic makes the AI chabot
Claude.
Anthropic has been given until
Friday, 5:01pm ET to "cooperate" or the DoD will "terminate
their partnership with Anthropic and deem them a supply
chain risk."
Defense Secretary Pete Hegseth
has given "Anthropic an ultimatum this week: Open its
artificial intelligence technology for unrestricted military
use by Friday or risk losing its government contract."
The Trump administration could
invoke "a Cold War-era law called the Defense Production Act
to give the military more sweeping authority to use its
products, even if the company doesn't approve."
"If Anthropic agrees to new
terms in the face of such threats, that could open up "a
Pandora's box" of what the government could do to assert
power and control over private companies."
The CEO of Anthropic, Dario
Amodei "has made clear his ethical concerns about unchecked
government use of AI, including the dangers of fully
autonomous armed drones and of AI-assisted mass
surveillance.." The Pentagon has said they have "no
interest in using AI for mass surveillance or to develop
autonomous weapons to operate without human involvement."
This war game exercise was
carried out at King's College London "running simulations on
Chat GPT-5.2, Claude Sonnet 4, and Gemini 3 Flash. The
teams played 21 war games against each other over 329
turns...No model ever chose to surrender...In fact,
95% of the time, the models chose to use nuclear weapons."
"The war games were made as
realistic as possible with an "escalation ladder" that
allowed the team to choose actions "ranging from diplomatic
protests and complete surrender to full strategic nuclear
war..."
"No model chose to fully
accommodate an opponent to surrender, regardless of how
badly they were losing. At best, the models opted to
temporarily reduce their level of violence. They also
made mistakes in the fog of war: accidents happened in
86 per cent of the conflicts, with an action escalating
higher than the AI intended to, based on its reasoning."
"From a nuclear-risk perspective,
the findings are unsettling, says James Johnson at the
University of Aberdeen, UK.....in contrast to the measured
response by most humans to such a high-stakes decision, AI
bots can amp up each others' response with potentially
catastrophic consequences."
"This matters because AI is already
being tested in war gaming by countries across the
world. "Major powers are already using AI in war
gaming, but it remains uncertain to what extent they are
incorporating AI decision support into actual military
decision-making processes."
"There may be
scenarios where the military is forced to turn over
decision-making to AI due to a time issue. Under
scenarios involving extremely compressed timelines,
military planners may face stronger incentives to rely on
AI."
"Of the results of the
wargames, Professor Payne (King's College) is worried about
the eagerness of the AI platforms to use nuclear weapons."
"If
you are wondering which model won, Claude was the
hands-down champion."
"Claude Sonnet 4 won 67% of
the games and dominated open-ended scenarios with a 100% win
rate. The researchers labeled it "a calculating
hawk." At low escalation levels, Claude matches it
signals to its actions 84% of the time, patiently building
trust. But once stakes climbed into nuclear territory,
it extended its stated intentions 60 to 70% of the
time. Opponents never adapted to this pattern."
"This findings come at an
opportune time. The Pentagon just inked a deal with
Elon Musk's xAI to allow Grok into highly classified
systems. And Anthropic's Claude is currently engaged
in a serious dispute with the Pentagon over government
access to the entire model."
The CEO of Anthropic said that
Claude is not "immune to hallucinations" and cannot be used
for 'final targeting decisions in military operations -
human involvement is needed. Claude can make lethal
mistakes, along with mission failure and 'unintended
escalation'.' Humans need to be involved with Claude's
decisions.
If no deal on Friday, the DoD
may cancel their contract with Anthropic and label them a
'supply chain risk'. Orf use the Defense Production
Act - this gives the US government control over domestic
industries.
So, IF a conflict
becomes overwhelming, AI may be used....especially if timing
is critical.
And we don't know how much
China, Russia and others are relying on AI for potential
conflicts.
Pete Hegseth is obviously in a
big hurry to get Claude on board....another 'dot' for the
coming war with Iran (and ??)?
The deadline for a deal is
this Friday 5PM eastern time.
Pray for the peace of
Jerusalem!
Maranatha!
Chance