So here I was, one day in the past and I see a BBC article. I saw the headline, I saw the ‘bully approach’ and initially I ignored it. It was not the BBC, there was no setting that seemingly truly interested me. I was thinking of a few settings towards IP that could give Apple (and optionally Meta) a nice boost. As I was mulling over the ideas I was having, in comes the CBC about 10 hours ago, or better stated I noticed their article and now something clicks in my mind. I started rereading the two articles. The BBC (at https://www.bbc.com/news/articles/cn48jj3y8ezo) gives us ‘Trump orders government to stop using Anthropic in battle over AI use’ with ““We don’t need it, we don’t want it, and will not do business with them again!” Trump wrote in a Truth Social post on Friday.” Of course if he doesn’t want it, there must be a good reason why people might want to use it and we are given “Anthropic is mired in a row with the White House after refusing demands that it agree to give the US military unfettered access to its AI tools. The refusal led US Defence Secretary Pete Hegseth to say he’s deemed Anthropic a “supply chain risk”.” And we are given the quandary that there should be some clarity. The idea that the US Military has unrestrained or uninhibited access to any AI is dangerous. And that is merely to look at it from THEIR point of view. We saw over the last 5 years a few examples where Pentagon staff used whatever USB key they had optionally opening their systems to backdoors and this can result in several ways where the Pentagon would be affected including: Human Interface Device (HID) Spoofing, Malware Infection via Social Engineering, Exploiting OS Vulnerabilities or Juice Jacking (Compromised Public Ports/Cables) and a few other ways. Even in this decade more than one system seemingly ended up on the danger list. So, ‘someone’ now wants to grant AI unfettered access which opens the doors to AI accessing data involves sophisticated, automated, and often, continuous interaction between intelligent systems and vast data sources, including internal corporate databases, cloud storage, and public web content. It constitutes a critical, high-speed, and high-stakes component of the modern AI ecosystem that raises significant security and privacy challenges. And this is not some ‘fear mongering’ There is a lot of AI works that is still to be considered and because AI doesn’t exist and this is all DML on several layers that interact there are dangers to be seen. As we saw a mere week ago that Microsoft had to ‘confess’ that it had accessed confidential emails of Microsoft users. Now consider this happening on a serious level in the Pentagon. It has well over 50,000 desktop computers within its building, with reports from 2014 indicating at least 18,000 were part of specific virtualized infrastructure. Now consider that we have seen the accusation of “Based on reports in early 2025 and 2026, OpenAI has accused Chinese AI startup DeepSeek of “inappropriately” distilling, or copying, the capabilities of OpenAI’s models (specifically ChatGPT and its reasoning models like o1) to train its own competing, low-cost models (such as DeepSeek-R1)”. As such, the dangers of unfettered access can go in two directions and that sets the bar of distilling from the Pentagon a lot lower than anyone could find acceptable. As such there is every chance that Russia is already considering the massive win they could gain once the unfettered access could merely hit one system that was transgressed upon. Because the greedy and the stupid will do anything to propel the setting of self, whilst not caring what others could gain in that setting as well.
So whilst some will consider the dangers of “The company said that “designating Anthropic as a supply chain risk would be an unprecedented action — one historically reserved for US adversaries, never before publicly applied to an American company.” Anthropic said the “designation would both be legally unsound and set a dangerous precedent for any American company that negotiates with the government.”” No one seems to be considering that the opposite is a lot more dangerous. So whilst some focus on the stage of “Anthropic had said it sought narrow assurances from the Pentagon that its AI chatbot Claude would not be used for mass surveillance of Americans or in fully autonomous weapons. The Pentagon said it was not interested in such uses and would only deploy the technology in legal ways, but it also insisted on access without any limitations. The government’s effort to assert dominance over the internal decision-making of the company comes amid a wider clash over AI’s role in national security and concerns about how increasingly capable machines could be used in high-stakes situations involving lethal force, sensitive information or government surveillance. Trump said Anthropic made a mistake trying to strong-arm the Pentagon. He wrote on Truth Social that most agencies must immediately stop using Anthropic’s AI but gave the Pentagon a six-month period to phase out the technology that is already embedded in military platforms.” As I personally see it, it is the accumulation of stupid and technologically ignorant all combined in one package. And that is before we get to mass surveillance. You see combine mass surveillance with data distilling and the United States of America will be handing the data on 349 million Americans straight to China and Russia. This is not AI, this is DML. That means it comes with the hangups and limitations of a programmer. So when this goes wrong it goes wrong in a massive way.
As such what will people like President Trump and Pete Hegseth say? Do they think that the response ‘Oops’ will cover it?
I don’t think so, but what do I know?
So whilst CBC (at https://www.cbc.ca/news/business/trump-anthropic-feud-ai-9.7109006) gives us “U.S. President Donald Trump, U.S. Defence Secretary Pete Hegseth and other officials took to social media to chastise Anthropic for failing to allow the military unrestricted use of its AI technology by a Friday deadline, accusing it of endangering national security after CEO Dario Amodei refused to back down over concerns the company’s products could be used in ways that would violate its safeguards.” And this is the setting we expect to see and it will be the undoing of several people, because as I see it “U.S. President Donald Trump, U.S. Defence Secretary Pete Hegseth and other officials” is the start of what comes next. You see, the internet doesn’t forget and these ‘other officials’ have sealed their fate with this action and there is no ‘He told me to do that’ they were instrumental in assisting to hand over the data of the population of the United States of America to optionally both China and Russia. Do you feel safe now?
And in response to this setting we see “The dispute stunned AI developers in Silicon Valley, where venture capitalists, prominent AI scientists and a large number of workers from Anthropic’s top rivals — OpenAI and Google — voiced support for Amodei’s stand in open letters and other forums.” And that should have been a clear message that the competition was on the side of Amodei, so, why would that be? Whilst people in the Pentagon (seemingly) forgot about that router with password ‘Cisco123’ there is every chance that these DML engines will be cleverly distilled by people controlling systems like DeepSeek and whatever the Russians have. I should buy another egg timer, because this is a setting that might gain me a few coins, especially as several people are blind to the danger that is coming for them. And consider one additional setting. It is said that:
“AI decisions can be corrupted primarily through the manipulation of the training data, the algorithms, or the inputs during operation. These methods can introduce biases, cause malfunctions, or lead to fraudulent outcomes, often making the corruption difficult to detect due to the complex “black box” nature of many AI systems.”
So what happens when distilling comes with an additional insertion of data? I can’t wait for that setting to lose balance and the training data in American data centers start losing authentication and reliability markers. But that is likely a story for another day.
Have a great day today.
Lawrence van Rijn - Law Lord to be 