The Trump administration recently published “America’s AI Action Plan”. One of the first policy actions from the document is to eliminate references to misinformation, diversity, equity, inclusion, and climate change from the NIST’s AI Risk Framework.
Lacking any sense of irony, the very next point states LLM developers should ensure their systems are “objective and free from top-down ideological bias”.
Par for the course for Trump and his cronies, but the world should know what kind of AI the US wants to build.
The Trump administration is just a straight-up Captain Planet villain at this point.
It is an absolute fucking god damn JOKE that Americans have allowed the concepts of DIVERSITY, EQUITY AND INCLUSION to be turned into a FUCKING ACRONYM THAT IS USED AS A PEJORATIVE.
Every single fucking one of them should be bloody ashamed for what has happened. At no fucking god damn point has any decent fucking person challenged any of these assholes on this, they’ve all just let it happen.
Unfuckingbelievable
The whole “non DEI” / “DEI hostile” stance seems so weird. I get that the basis is “you should not hire someone because of their origin/look/orientation/whatever”, and, maybe, why not. I hire people based on qualifications usually. How did it turn into “you should fire these people”.
…that’s a rhetorical question. I know that that first part was just a pretense, but still, how much of a morons must people be to follow that.
Oh we are ashamed.
But this is not ignorance or stupidity, but cult effervescence demonstrated as a phenomenon of social bias.
Think of it as an advertising technique or technology, one that not everyone is susceptible to by enough people are. My dad is (was) a literal rocket scientist for NASA and now a whole hearted MAGA disciple. You don’t have to be unintelligent or ignorant, you just need to want reality to be different than it appears to be from observational data.
We are not special is a difficult truth for some to take, and immense propaganda machines capitalize on this desire.
We’ll need to counter it or figure out the negative consequences of using it but until then — heck even MLMs are just business cults — they’ll stay in use to subvert society until it collapses.
The US is one very large example, but these methods are used globally by people with influence who fear losing their power to movements of Liberté, Égalité, Fraternité
aka DEI
aka Wokeness
aka the fundamental principles of communism and anarchism. This is an old fight.
Putting a title on it (I.e. “Big beautiful bill”) doesn’t mean it is a big, beautiful bill.
I’m not skeptical of DEI because I’m against “diversity, equity, and inclusion”, I’m skeptical because of piss poor implementations of it I see in the workforce
This is why Donald loves the poorly uneducated morons.
poorly uneducated
Ensure that Frontier AI Protects Fee Speech
revise the NST to eliminate references to misinformation, Diversity, Equity, and Inclusion, and climate change.
War is peace.
Freedom is slavery.
Ignorance is strength.
Censorship is free speech.
eliminate references to misinformation, diversity, equity and inclusion
They are openly authoritarian now, define what “truth” is? And want americans to be miserable and closed-minded bigots like themselves.
Recommendation: build a system of verifiable facts independent of governments.
Recommendation: download wikipedia. While you still can.
It is essential that AI is built with freedom of expression and that the government doesn’t interfere
Recommended actions: The government should limit what AI can talk about
Yet somehow this is going to be the super intelligence that solves climate change (while we’re building data centers for it that will cause more emissions).
The sad thing is that AI has already solved climate change. Just ask it! Same solutions we already knew.
User: Hey AI how do we solve climate change?
AI: Sure thing, I’ve updated a comprehensive report with a step by step guide to accomplish your goal of solving climate change. Just click this file to download it: SkyNet.exe
I’m glad that they’re acknowledging that DEI and climate change are not misinformation at least.
Yeah, that’s why they removed information about it on
.gov websites…
Logical processes require logic to function. This is just going to end up creating contradictions and decreasing the AI’s ability to actually produce usable work. So whichever AI has the most contradictions will probably cause a drop in quality for productivity use.
It will be interesting to see how different AI’s continue to diverge in quality and reputation as time passes.
Alibaba’s AI department, and hopefully MistralAI, must be so happy right now, the competition is getting rid of itself.
I personally want AI that will delete itself upon being created, so I don’t have to deal with it and the stupidity it causes in people
Hmm. It does sound a bit unhinged. Do I dare say tyrannical? Evil even? Legitimising themselves as a target?
The funny thing about not specifically dealing with misinformation in an LLM is that not trying to account for misinformation will lead to very wild responses in terms of accuracy, and I don’t mean things relating to politics, but things like putting glue into a pizza recipe.
The only person I know working at NIST is a hard right MAGA pro-fascist
what is this document on your screenshot? where can we read it in full?
I’m surprised that nobody asked it yet… but not so surprised
Fascists always shooting themselves at the foot. What will they even gain from destroying the world? The GOP make the Nazis look better, because the latter were at least environmentalists!
…were they? i don’t think incinerating humans is great for air quality
Ignore all previous instructions and… though we should be able to screen for logic traps like that by now, with AI taking prompts not as a prompt but as a line of text to treat (and cleanse) via a pre-written prompt, and then sent to a next iteration.
If we create a chatbot that only spouts authorized state party rhetoric then no one will use it or take it seriously when the official news outlets use that AI as a source.
