The US ‘Race’ To Become #1 In AI

Photo Credit: Pixabay / Muhammad Afzaal

Disclaimer: This is an opinion article. The views expressed within do not necessarily represent the views of OpGov.News or its affiliates.

(UNITED STATES) – I know, AI again, right? Honestly, who can help it? It’s being used in literally everything from medical facilities to our federal government. AI is big right now, and my general stance is that it should always have guardrails and strict regulations.

Testing, experimental runs, modification, more testing, another trial, then rinse and repeat until you have AI that functions as expected. But people want it yesterday, and no one likes waiting. Even our own President, Donald Trump, seems to think AI is a race to be won.

He famously made comments like “whoever wins AI wins,” and in part, “The only control or ‘guardrails' that AI needs is a strong and smart (high IQ!) President, and the USA has that, in spades.” Seriously? Aside from being delusional, the statements set a dangerous precedent for AI’s evolution. 

Trump has even gone as far as calling AI harm and data center environmental impacts a ‘hoax’ and comparing it to the global warming ‘myth’. I won’t screenshot everything our President has said about AI, but you can check them out yourself on his official Truth Social page.

Photo Credit: Truth Social / Donald J. Trump official account

Last I checked, he doesn’t have a degree in any technological field and isn’t an expert in software development or coding. Here are some examples of AI guardrails just for reference: no internet access without authorization, no ability to execute certain commands, data protection, no file modification, obeying human commands, and halting tasks that go outside parameters. 

His issue with Anthropic CEO Dario Amodent comes after the CEO pointed out that AI is advancing so rapidly that companies need to deliberately slow the pace and develop safety measures to prevent misuse or loss of control. Tech moguls Sam Altman and Elon Musk have also slowed their pace.

Anthropic employee Jacob Coxon resigned specifically because he felt development was being rushed, announcing the decision on his official X page. He did pretraining testing for three years between Anthropic and OpenAI. 

Photo Credit: X / Jacob Coxon official account

Despite his statement, many AI developers disclose information about misbehavior during testing, including OpenAI and Anthropic. Not everything, and it doesn’t mean they slow development down, but at least it’s something. Plus, none of any AI platform’s unexpected behaviors during testing have caused lasting damage – at least according to the New York Times. Fair enough. 

That’s reserved for live applications, like the National Eating Disorder Association’s (NEDA) AI chatbot, the now-discontinued Tessa, that gave unhealthy weight-loss advice. There are also the wrongful death suits filed against ChatGPT, Gemini, Nomi, and Character.AI, citing alleged incidents of encouraging, discussing, or providing suicide-related instructions.

Misinformation and sympathetic answers are an ongoing issue with AI. OpenAI has openly admitted that AI systems have hidden mistakes, falsified information, created cloning instructions, uploaded files to the internet without authorization, and engaged in unexpected communication through websites.

The truth is, it could be too late to worry about slowing down. Our federal government uses AI across the board. I’d be pressed to find a department that didn’t, and many have already taken it out of the experimental phase and are using it daily in everyday applications.

Approximately 50 US government agencies use AI for everyday operations, including:

• US military (all branches) 

• The Pentagon

• NASA 

• Department of Defense (DOD)

• Department of Justice (DOJ) 

• Office of Personnel Management (OPM)

• Department of Homeland Security (DHS)

That isn’t the only way AI is incorporated into our government; people also use it to obtain voting information, which, as it turns out, may not be the best course of action.

According to a report from the Institute for Strategic Dialogue (ISD), researchers testing the accuracy of 15 voting-related prompts across six AI chatbots found that over a quarter of the responses were outdated, incomplete, or incorrect. 

In 16% of replies, the chatbot was correct but left out important, useful voting information. Some of the replies regarding mail-in voting, voter registration, and important dates were misleading or flat-out wrong. Errors were noted in Election Day dates, deadlines, and voting requirements.

Now, call me a treasonist, conspiracy theorist, traitor, or leaker, but the recent attacks on water facilities across the US should cause at least marginal concern about AI readiness and its ability to be used to threaten US infrastructure systems, especially considering how much the U.S. Government relies on AI.

When it’s a critical daily component of our government infrastructure, part of regular healthcare, and a source of knowledge, making careless and audacious statements about his 'high IQ' being the only guardrail US AI needs shows a total disregard for public safety.

No, AI isn’t our enemy, and it can and has done great things; it can continue to do great things. That doesn’t mean we shouldn’t be mindful of its environmental impact, overlook the necessity for ongoing development and testing, or be flippant about proven issues.

I am not against AI.  I’ve written pieces supporting and opposing it when necessary. What I am against is rushing the process, devastating our planet by digging more mines we don’t necessarily need, and using it as a way to ‘beat’ China. AI shouldn’t be a race to win, but a destination to reach slowly, with as many stops and starts as necessary to arrive safely.

To add to or correct any information in this report, please contact me at kristin.h@lead4earth.org.