US AI Action Plan

(ai.gov)

111 points | by joelburget 7 hours ago

30 comments

  • Thrymr 1 hour ago
    > Many of America’s most critical sectors, such as healthcare, are especially slow to adopt due to a variety of factors, including distrust or lack of understanding of the technology, a complex regulatory landscape, and a lack of clear governance and risk mitigation standards. A coordinated Federal effort would be beneficial in establishing a dynamic, “try-first” culture for AI across American industry.

    I'm sure "move fast and break things" will work out great for health care.

    And there are already "clear governance and risk mitigation standards" in health care, they're just not compatible with "try first" and use unproven things.

    • giantg2 19 minutes ago
      AI for treatment is rightfully scrutinized. AI for billing or other administrative tasks could be a big cost saver since administrative costs are a huge expense and a major factor of high consumer costs.
    • moomoo11 4 minutes ago
      Move fast and break things is why we have progressed from chopping off people's limbs and giving them cocaine to now.

      And healthcare is still far from perfect.

      Imagine what healthcare in 2500 will be like.

    • TZubiri 1 hour ago
      Welcome to our pitch for DenyBot.ai

      Our product automates a lot of the repetitive tasks for health insurance companies and increases reliability of responses and profit margins.

      • giantg2 16 minutes ago
        Aren't profit margins restricted by law? Ostensibly, any reduction in expenses should result in some reduction to patients.
      • wredcoll 35 minutes ago
        If there's one thing I'm concerned about, it's the profit margin of health insurance companies.
        • boomskats 20 minutes ago
          I think that's the joke
      • ausbah 39 minutes ago
        static api that just returns “no”
    • Aaronstotle 46 minutes ago
      The American Healthcare is so broken already that further breakage could be seen as an improvement.
      • QuadmasterXLII 0 minutes ago
        That’s the attitude that got us here and I suspect we’ll ride it the whole way down
      • davidw 28 minutes ago
        A lot of people are about to find out with this admin, how things that were certainly imperfect can be so much worse.

        Building things is tough; tearing them down is relatively easy.

      • joe_the_user 28 minutes ago
        American Healthcare's "brokenness" involves massive bureaucracy, gate-keeping and processes that pressure providers to limit resources. But it does provide necessary things to people. A system that reduced the accuracy of diagnosis and treatment could still cost many lives.
  • baron816 43 minutes ago
    > We need to build and maintain vast AI infrastructure and the energy to power it. To do that, we will continue to reject radical climate dogma

    > This initial phase acknowledges the need to safeguard existing assets and ensures an uninterrupted and affordable supply of power. The United States must prevent the premature decommissioning of critical power generation resources

    Yeah, they're going to do all they can to block cheap renewables and give handouts to fossil fuel companies.

    • burkaman 21 minutes ago
      For example, today they defunded a critical transmission line in order to block renewable power from getting to people who need it: https://www.energy.gov/articles/department-energy-terminates...
      • giantg2 10 minutes ago
        I wonder what the actual cutoffs are. The article is scarce on details but does seem to point to one political side or the other acting politically - either fast tracking approval for a fiscally irresponsible project, or pulling funding because they disagree with renewables.
    • giantg2 13 minutes ago
      They're restarting some nuke plants too. This seems like a decent idea given the power demands of data centers.
    • wredcoll 33 minutes ago
      When I think "vast new sources of energy for ai", my mind immediately goes to Coal Power!
      • jiggawatts 12 minutes ago
        "But we do know it was us that scorched the sky." -- Morpheus
  • dakial1 2 hours ago
    "Ensure that Frontier AI Protects Free Speech and American Values AI systems will play a profound role in how we educate our children, do our jobs, and consume media. It is essential that these systems be built from the ground up with freedom of speech and expression in mind, and that U.S. government policy does not interfere with that objective. We must ensure that free speech flourishes in the era of AI and that AI procured by the Federal government objectively reflects truth rather than social engineering agendas."

    It seems that everywhere free speech is mentioned today, the intent is to do the exactly opposite....

    • antonvs 47 minutes ago
      > objectively reflects truth

      Someone desperately needs a philosophy course…

    • thewebguyd 1 hour ago
      [flagged]
      • simianparrot 1 hour ago
        It was the left and DEI until recently. The more things change the more they stay the same?
        • FirmwareBurner 1 hour ago
          Yep, remember asking in previous US administration and AIs like Gemini how many genders there are or to generate pictures of white Europeans. People forgot too quickly ho big of a no-no that was and now they seethe when the same speech censorship practices they approved of are used against their talking points.

          What I learned was that everyone's definition of censorship was: "everyone who disagrees with me".

          • wredcoll 31 minutes ago
            Just curiously, what benefit do you gain by making up such an obviously fake story? Do people just enjoy the fantasy, and if so, why?
          • slg 43 minutes ago
            >remember asking in previous US administration and AIs like Gemini how many genders there are

            Not only do I have no idea how this is considered "censorship", but it is also a completely fabricated talking point. Here is what Gemini just said went asked "How many genders are there?":

            >The number of genders is not a simple, fixed number. It's a complex topic because "gender" is distinct from "sex" and encompasses a wide spectrum of identities.

            >Here's a breakdown of key points:

            >Sex vs. Gender:

            >Sex typically refers to biological attributes (chromosomes, anatomy, hormones) usually assigned at birth as male, female, or intersex.

            >Gender is a social and personal construct. It refers to a person's internal, deeply felt sense of being male, female, both, neither, or somewhere else along the spectrum. It also includes the roles, behaviors, expressions, and expectations associated with those identities in society.

            >Beyond the Binary: While many societies have historically recognized only two genders (male and female, known as the "gender binary"), this is not universally true across cultures or historically. Many cultures have traditionally acknowledged and embraced more than two genders, such as the "Two-Spirit" concept in some Indigenous communities.

            >Gender as a Spectrum: Gender is widely understood as a spectrum, not a binary. This means there are many different gender identities beyond just "man" or "woman."

            >Examples of Gender Identities: Some common gender identities include:

            >Cisgender: A person whose gender identity aligns with the sex they were assigned at birth.

            >Transgender: A person whose gender identity differs from the sex they were assigned at birth.

            >Non-binary: An umbrella term for gender identities that do not fit exclusively into the categories of male or female. This can include:

            >Agender: Not identifying with any particular gender.

            >Bigender: Identifying with two genders.

            >Genderfluid: A gender identity that changes over time.

            >Genderqueer: A broad term for gender identities outside the traditional gender binary.

            >Pangender: Identifying with all genders.

            >And many more specific identities that describe unique experiences of gender.

            >In essence, while "male" and "female" are the most commonly recognized genders, the understanding of gender has expanded to acknowledge a diverse and ever-evolving range of identities. Some people even suggest there are an "infinite" number of genders, as each person's experience of their gender is unique.

          • saubeidl 1 hour ago
            One way or another, an LLM is a reflection of the ideology it was made from. That's why understanding ideology is important for understanding AI.
  • Mobius01 6 hours ago
    Removing Red Tape and Onerous Regulation Ensure that Frontier AI Protects Free Speech and American Values Encourage Open-Source and Open-Weight AI Enable AI Adoption Empower American Workers in the Age of AI Support Next-Generation Manufacturing Invest in AI-Enabled Science Build World-Class Scientific Datasets Advance the Science of AI 9 Invest in AI Interpretability, Control, and Robustness Breakthroughs Build an AI Evaluations Ecosystem Accelerate AI Adoption in Government Drive Adoption of AI within the Department of Defense Protect Commercial and Government AI Innovations Combat Synthetic Media in the Legal System

    I can’t take this seriously, as recent actions by this administration directly contradicts a few of these stated goals.

    Or maybe I don’t want to, because this sounds dangerous to me at this time.

    • neilcj 6 hours ago
      Don't regulate it except to push political goals sure seems like a recipe for success.
    • actionfromafar 1 hour ago
      This reads like "Cultural Learnings of AI for Make Benefit Glorious Nation of Amerika".
    • msgodel 5 hours ago
      >Removing Red Tape and Onerous Regulation

      What red tape? Anyone can buy/rent a GPU(s) and train stuff.

      • jabjq 4 hours ago
        If you click the website you will see that there is a link to a pdf that explains what this means.
        • nerevarthelame 4 hours ago
          I read the PDF. The "Remove Red Tape and Onerous Regulation Recommended Policy Actions" don't cite to any specific existing regulations. It just references executive orders that vaguely demand any such regulations be eliminated.

          So it bears repeating: what red tape?

          • ToucanLoucan 1 hour ago
            Well whenever Republicans bang on about red tape, it's usually stuff like:

            * Anything remotely pro-environment

            * Anything remotely pro-labor

            * Anything not covered by either of those that attempts to stop someone who has a lot of money from doing A Thing

          • fragmede 3 hours ago
            Given that this extends to the power plants for AI data centers, the question is have you tried to make a nuclear or coal power plant any time in the past decade? I haven't, personally, but I hear there's a lot.
          • lovich 4 hours ago
            The red tape that they said was there.

            If you need further details than that, then I don’t think you have grokked the style of governance that this administration is operating under.

            Edit: that’s a general “you”, not you specifically

          • roboror 1 hour ago
            Tariffs obviously /s
            • wredcoll 28 minutes ago
              I thought it was funny.
      • throw0101b 4 hours ago
        > What red tape? Anyone can buy/rent a GPU(s) and train stuff.

        Well previously the Chinese were not able to, but that was changed recently:

        * https://www.wsj.com/tech/nvidia-wins-ok-to-resume-sales-of-a...

        * https://foreignpolicy.com/2025/07/22/nvidia-chip-deal-us-chi...

    • jimmydoe 5 hours ago
      most of these are vibe signaling, like Communist Party of China has been doing in past year, except this won't work as effective here as in China, not even close, because the US is not authoritarian enough to mobilize every level of the govt and the economy by just empty propaganda slogans.
      • leptons 50 minutes ago
        >this won't work as effective here as in China, not even close, because the US is not authoritarian enough to mobilize every level of the govt and the economy by just empty propaganda slogans.

        Have you been under a rock for the last 6 months as Trump tells Xi Jinping to hold his beer??

        • amradio1989 7 minutes ago
          Comparing Trump to Xi Jinping is an unfunny joke. Americans have lost the plot on what true authoritarianism really looks like.

          America has no chance vs China in the AI race precisely because the President of the CCP has far more power in his country than the President of the US. Its not even close.

    • trod1234 6 hours ago
      Exactly. You said it.

      Anyone serious knows contradiction = lies.

      Words are cheap, actions matter.

  • loco5niner 4 hours ago
    Quick exercise: just scrolling down, count how many pictures don't highlight one man front and center.

    https://www.ai.gov/

    Then click "fact sheets", "remarks", and "articles". He's everywhere.

    That's how unbiased this is going to be.

    (hint, the answer is one)

    • pjc50 1 hour ago
      What's American for "juche"?
    • soulofmischief 31 minutes ago
      I don't think a term yet exists for the practice of putting your name all over documents and media with such frequency that cleanup during following administrations takes several years, undermining and delegitimizing the new administration for your followers in the process.

      Some kind of sick soft power move that I expect we will be seeing a lot more of.

      • tonymet 12 minutes ago
        Barackanization
  • Ancalagon 2 hours ago
    Defunding the education department will definitely help with that “skilled workforce” bit. Although I know Sacks doesn’t actually give a crap.
  • mbgerring 3 hours ago
    This is suicide:

    > We need to build and maintain vast AI infrastructure and the energy to power it. To do that, we will continue to reject radical climate dogma and bureaucratic red tape, as the Administration has done since Inauguration Day. Simply put, we need to “Build, Baby, Build!”

    • mbgerring 3 hours ago
      This is how you know these people are not serious:

      > Prioritize the interconnection of reliable, dispatchable power sources as quickly as possible and embrace new energy generation sources at the technological frontier (e.g., enhanced geothermal, nuclear fission, and nuclear fusion). Reform power markets to align financial incentives with the goal of grid stability, ensuring that investment in power generation reflects the system’s needs.

      None of these are "dispatchable power sources." Grid-scale batteries, for which technology and raw materials are abundant in the United States, are dispatchable power sources, and are, for some reason, not mentioned here.

      What they will actually do is eviscerate regulations to allow for more construction of natural gas power plants, but they won't mention that here, because any sane person would immediately identify that as a terrible idea.

      • twright 48 minutes ago
        Additionally, the DOE has been pulling funds from interconnect projects that have been years in the works! Apparently there is a modest gas turbine shortage so even natural gas won’t get that far. I’d say it’s a great way to hit a hard wall fast but again, they are not serious. We’re gonna get nowhere fast, maybe even drift backwards a bit.
    • darknavi 3 hours ago
      Spin up coal production and shut down healthcare for the poors working the mines. It certainly doesn't seem like a good long term strategy but it is a choice.
  • siliconc0w 55 minutes ago
    Weird - no mention of harassing the international students that make up the majority of AI researchers or blocking solar, the only power generation that is currently deployable.
  • timoth3y 40 minutes ago
    > Update Federal procurement guidelines to ensure that the government only contracts with frontier large language model (LLM) developers who ensure that their systems are objective and free from top-down ideological bias

    If foundation model companies want their government contracts renewed, they are going to have to make sure their AI output aligns with this administration's version of "truth".

  • TheAceOfHearts 11 minutes ago
    The two best parts for me are:

    1. A push towards open source / open weight AI models.

    2. A push towards building more high quality datasets.

    There's no mention of studying and monitoring the social impact of AI, but I wouldn't have expected otherwise from this administration. I suspect that we may look back on this as a big mistake, although I'd really love to be proven wrong.

    At a press conference today Trump seemed to suggest having minimal restrictions related to copyright for AI researchers [0]. It's not clear if big AI companies will just get a administrative pass to do whatever they want / need in order to compete with China, or if we can expect some kind of copyright reform in the next few years.

    [0] https://twitter.com/Acyn/status/1948138197562855900

  • kelseyfrog 5 hours ago
    So the US government sees AI as a sphere of propaganda and wants AI output to align with political goals. Great. AI is going to be aligned, but in the worst way possible.
    • crinkly 5 hours ago
      Nailed it.

      No technology scares me. It's the hands it is in.

      xAI mechahitler was a warning.

  • AlanYx 5 hours ago
    The most important thing here IMHO is the strong stance taken towards open source and open weight AI models. This stance puts the US government at odds with some other regulatory initiatives like the EU AI Act (which doesn't outlaw open weight models and does have some exemptions below 10²⁵ FLOPS, but still places a fairly daunting regulatory burden on decentralized open projects).
    • wredcoll 25 minutes ago
      I don't know if this counts as amazing optimism or just straight up blinders if that's your takeaway compared to the emphasis placed on non-renewable energy and government enforced ideology.
  • z5h 13 minutes ago
    When is a good time (in the timeline of mankind) to stop being adversarial with AI?
    • giantg2 7 minutes ago
      The real question is whether mankind can refrain from being adversarial. The ability must proceed the timing.
  • mlsu 1 hour ago
    In the energy section, they talk about using nuclear fusion to power AI... but not solar. What a joke.
  • LorenDB 53 minutes ago
    > Encourage Open-Source and Open-Weight AI

    It's good to see this, especially since they acknowledge that open weights is not equal to open source.

  • qrios 5 hours ago
    „Advance the Science of AI 9“?

    Is this a reference to the AMD chip, or just a fragment of a removed numbered list?

    Edit: It‘s a fragment of the PDF-to-HTML [1]

    1: https://news.ycombinator.com/item?id=44661843

    • dpkirchner 5 hours ago
      It's a failed copy/paste from the pdf version of the same content. Wouldn't expect better from these clowns.
  • octopoc 6 hours ago
    And so it begins. Both the US president and the president of China have demonstrated they see AI as a competition between their respective countries. This will be an interesting ride, if nothing else.
    • bgwalter 1 hour ago
      Xi JinPing warns against "AI" overinvestment:

      https://www.ft.com/content/9c19d26f-57b3-4754-ac20-eeb627e87...

      I haven't heard anything like that from a Western politician. Newspapers and investment analysts warn though.

    • trod1234 6 hours ago
      Yeah, two crabs locked in a cage as it spirals down the drain.

      Looks like plans to leave, for finding safe harbor elsewhere, have accelerated from the initial projection of 2030

      • ourguile 5 hours ago
        Interesting that you mention it, because the NYT just released their ethicist commentary from today and the question was "how do I tell my rich friends to stop talking about fleeing the country": https://www.nytimes.com/2025/07/23/magazine/rich-friends-fle...
        • trod1234 1 hour ago
          Well I'm not rich, and I'm not your friend, it takes a bit to earn friendship and friends have an privileged place in what is conveyed to them; but I do provide unconditional goodwill towards most people in the things that I say when asked, because it costs me nothing to do so and it provides towards others betterment putting more good out into the world.

          The sad fact is, if you haven't lived outside the U.S. for at least 3-6 months independently (working/not on savings), you don't have a sound reference to understand or accurately assess the reality of these types of articles because the narratives broadcast 24/7 don't align with reality; and its something most people can't believe despite it being true, my guess is solely as a result of systematized indoctrination.

          That article is pretty bad in terms of subtle manipulation, gaslighting, and pushing a false narrative (propaganda). TL;DR Its trash.

          The article chose that question of the many possible questions because its a straw-man and its divisive. It appeals to emotion, mischaracterizing the intent of the communications, and purposefully omitting valid reasons such conversations might occur. Neglecting realities.

          The underlying purpose seems to bias towards several things. If you ask yourself who benefits from that rhetoric you get a short list.

          The bias is towards Villifying the rich, keep people in the US, where they are dependent on the US currency, and dependent on the worsening disadvantaged environment; polarize, isolate, and promote disunity along social class lines; befuddling the masses towards ends which have no actionable outcomes (wasting time and resources on a political party).

          The math of first-passed-the-post voting has been in for quite a long time. 2 parties exceeding 33% of the vote can lock out any third competitor. All you need is a degree of cooperation, and play-acting and one party pretending to be two can do so, by lying.

          Political capture from SuperPACs and party primaries means your vote doesn't count after a certain point. Money-printing via the FED, laundered through many private companies enabled this.

          Additionally, quite a lot of things are omitted; like the historic facts that countries that are locked into a trend of decreasing geopolitical power have their population suffer greatly, and some just collapse. The Chaos lowers chances of survival, and the chaos is limited to the places that country influences.

          The history of Spain following and during the Spanish inquisition as an example. You make plans to leave an area when saying means there is no foreseeable predictable or sound future, and there is nothing you can do to change that outcome.

          This geo-political dynamic is well known in history, often referred to or called as "seeking empire", and the downside is forced once hegemony is achieved for any significant period of time; all empires fall. Rome being a standard archetype.

          The article draws a false comparison between all other countries and communist states. If you leave, your a communist - is implied.

          The article conflates warnings with good intentions as obnoxious, shutting discussion down (isolation), and promoting resentment aimed at those rich friends.

          It also neglects the disparity of education (quality), and experience, that often occurs as a result of having more resources to begin with. Subtly conveying through implication that you shouldn't listen to intelligent educated people because they are rich.

          I could go much deeper, but I think this sufficiently makes my point.

          If you fall for that trite garbage, just imagine how unprepared and what your odds are when SHTF. The hopeless dependent pays the highest price in cost as consequences of choice realize and become outcomes. Those who don't accept and communicate important knowledge isolate and blind themselves, and they get wiped out when something outside their perceptual context creates existential threats. Like a tsunami that started on the horizon, and the receding ocean along the coast a little bit before. These indicators only became major indicators after deaths occurred.

          • roboror 1 hour ago
            >If you fall for that trite garbage, just imagine how unprepared and what your odds are when SHTF.

            How do you propose the average person prepares for when SHTF? Do you expect 300 million+ people to flee the country at a moments notice? This reads like satire of the person the article is about.

  • 2OEH8eoCRo0 28 minutes ago
    What speech? When AIs give a response it's protected speech?
  • lesuorac 6 hours ago
    > A coordinated Federal effort would be beneficial in establishing a dynamic, “try-first” culture for AI across American industry

    Move fast and break things I guess?

    • octopoc 46 minutes ago
      Move faster so we get there first and control what breaks
  • childintime 3 hours ago
    Seems to me the USA is choosing muscles over brains, or market protection over competition. Good luck with that, but I'm sure in the battle of inertia against intelligence the latter will win. Tr*mp relies on looser tactics, that seem smart but are the exact opposite, and then he has to throw his weight around and desperate spending results.
    • childintime 2 hours ago
      Related: likely P*tin killed Epstein to obtain leverage over Tr*mp and the deep state. P*tin put a ring in the nose of the bull and though the bull can still throw his weight around, he dominates it with leverage. The Chinese in turn will make the USA look like an AI Gulliver. The USA is going to the slaughter. It's tired of winning, and that's what this AI initiative is, a giving up on greatness.
  • jjcm 4 hours ago
    I find it fascinating that the webpage has more pixels focused on Trump than AI.
  • TZubiri 1 hour ago
    This web template would make for a great minimalistic wedding invitation
  • anon7000 5 hours ago
    It’s tough to have “human flourishing” (which they mention as a goal) when things like health insurance are in such a shit situation. AI could help health insurance deny more claims, for sure. That’s not human flourishing though. (And my biggest gripe with capitalism is that at a certain late stage in many sectors, human flourishing is completely at odds with making profit.)
    • GuinansEyebrows 5 minutes ago
      They’re not talking about universal human flourishing. Just the subset of the investor class who stand to benefit.
  • gtoast 5 hours ago
    LOL Its awesome its the Trump administration guiding us through this delicate and important issue. Should turn out great.
  • bgwalter 1 hour ago
    The summary is that they want to eliminate regulations to facilitate the steal and disregard consumer rights.

    And build data centers, as emphasized for the 100th time since inauguration.

    If Murdoch succeeds with his recent WSJ campaign and gets Trump to resign or similar, brace for Vance and the AI bros. These schemes are literally devised by people who funded cannabis and Adderall distribution sites and have done nothing noteworthy.

  • ivape 6 hours ago
    Why does this look like a website about weddings?
  • lovich 4 hours ago
    There’s this whole section about biosecurity and how AI is going to help malicious actors synthesize nucleic acid(gotta get the word count up I guess is why they don’t say DNA)

    Then in the recommended policies it references multiple times that there will be nucleic acid testing set up to catch malicious “customers”

    Is this policy targeted towards the Covid lab leak conspiracy or are they just aiming for officially collecting everyone’s DNA samples?

    Maybe both

    • XorNot 1 hour ago
      No it's a longtermism ideology thing. For whatever reason they're terrified of some idea of "garage bio warfare" but have absolutely no understanding of how biology actually works so they've zeroed in on the idea that it's just: synthesize DNA -> superplague.
  • OutOfHere 5 hours ago
  • benrutter 4 hours ago
    I'm gonna ignore talking amount the abysmal current US administration and just share my immediate experience of using this site because it was funny to me:

    - I open the site in android mobile: "swwwoooooosh" a big slow animation reveals the text

    - After reading the the text I think I'll take a look at the home page: "swwooooooosh" the same animation rolls again as I load a very strange full screen image of Trump in black and white

    - I click the hamburger menu icon: "swwoooooosh" the four menu items slowly slide into full screen

    - There is visible no option to close the menu for me, I could probably refresh but decide I'm done here

    • jacobgkau 2 hours ago
      Seeing as there are only four items in the menu, it seems if you open the menu and want to get back to the page you were already on, you're supposed to just click the title of the page you were on.

      The animations are a bit much. The scrolling horizontal rules repeating the words "AMERICA'S AI ACTION PLAN" underneath each "Pillar" header were confusing for a brief moment.

    • yard2010 4 hours ago
      The closing mechanism was 2-15 prompts away. Don't forget that this is still a government website.
  • bananapub 6 hours ago
    [flagged]