As you all know there are many different versions of “AI” from Expert Systems to “LLM Chatboxes”, e.g. ChatGPT. It’s the Deep Learning chatboxes that are of concern.
Joe Biden’s Exec Order 14110 outlined his policy goals re. AI development.
a) promoting competition in the AI industry
b) preventing AI-enabled threats to civil liberties
d) ensuring U.S. global competitiveness in AI
e) requiring a number of major federal agencies to create dedicated “chief artificial intelligence officer” positions
Biden set out 8 guiding principles such as “good for USA”, “good for Americans” etc but also stuff like:
“appropriate safeguards against fraud, unintended bias, discrimination, infringements on privacy & other harms from AI”;
“privacy & civil liberties must be protected”;
“collection, use & retention of data is lawful, secure & mitigates privacy & confidentiality risks. Agencies shall use available policy and technical tools, including privacy-enhancing technologies (PETs) to protect privacy & to combat the broader legal & societal risks that result from the improper collection & use of people’s data.“;
“engage with international allies & partners in developing a framework to manage AI’s risks, unlock AI’s potential for good & promote common approaches to shared challenges.”;
“will seek to promote responsible AI safety & security principles & actions with other nations, including competitors, while leading key global conversations & collaborations to ensure that AI benefits the whole world, rather than exacerbating inequities, threatening human rights & causing other harms.“
To the Tech Bros’ delight, Trump immediately moved to cancel 14110 on Day 1 (Jan 20th) of assuming office. He must have felt it was important to do that.
More recently Trump reminded the world that what he’d done by removing all constraint on AI product developers would give them a free hand to design the sort that of products that they, Trump and MAGA were expecting and that this would quickly “remove ‘woke’ from the world”.
In the light of the above, now consider the below…
The “Tech Bros” are watching you… and Trump and Project 2025 are watching them.
Roundabout 10th July, 2025, xAI, Musk’s AI company, gave us a glimpse inside Pandora’s Box.
Musk suddenly decided that Grok, xAI’s chatbox, had become “woke” and also was being critical of him.
So its program was altered to fix that – but then it went antisemitic and pro-Hitler. So xAI adjusted it again.
Now we know AI is highly sensitive, not only to what training material is selected to be fed to it, but also, in response to the weightings written into the algorithms. In this way, the output of any chatbox can be biased to reflect the morals/prejudices of its designer or owner or controller.
In other words we have had a demonstration as to how easily AI and chatbox output could be deliberately set up to promote the views and the interests of its owners – or that of the government or that of a particular creed or political dogma.
Once established widely you will be fed consistently biased output – with any alternative either impossible or impractical or difficult to access.
Your thoughts will become moulded.
Your thoughts will become their thoughts.
No more yellow umbrellas.
********************************************
This concept is rehearsed in George Orwell’s “1984”.
Those persons in control of the development of “Newspeak”, the Party’s engineered language, were deliberately designing it to narrow down thought by drastically reducing vocabulary and eliminating synonyms and complex structures.
By controlling language, the Party aimed to control thought, making “thoughtcrime” impossible because people will lack the words to express dissenting ideas. The novel argues that language is fundamental to thought, and by simplifying language, the Party can simplify thought and control reality itself.