On July 23, the Trump administration released its long -awaited AI Action Plan. The lack of copyright discounts for model training, the administration appears to be ready to give everything to the White House asking the White House during public consultation. However, according to Travis Hall, director of state engagement at the Center for Democracy and Technology, Trump’s policy vision will put states and tech companies themselves into “extraordinary regular uncertainty”.
It begins with an attempt to prevent Trump’s states from organizing the AI system. In the original draft of his recently approved tax megabil, the president has also included an amendment in which any state -level AI regulation will be imposed for 10 years. Finally, the clause was removed from the Senate in a decisive 99-1 vote.
It seems that Trump has not received a message. In his action plan, the president has indicated that he will order federal agencies to fund the “AI -related” states without the “burdensome” AI rules.
“It is not really clear which discretionary funds will be considered ‘related to AI’, and it is unclear what existing state rules-and which future suggestions-which will be considered ‘burdensome’ or as a” barrier “for the utility of federal funds. This is the design of the hall,” and it is a source. “
With this proposal, almost any discretionary fund can be considered as AI. Hall suggests a scenario where a law like Colorado Artificial Intelligence Act (CAIA), which is designed to protect people from algorithmic discrimination, can be seen as a barrier to funds for providing technology enrichment to schools because they are seeing a hindrance to their students.
The possibility of reading “AI related” is far -reaching. Everything from broadband to the Highway Infrastructure Fund can be at risk as machine learning technologies have begun to touch every part of modern life.
Its itself, it will be bad, but the president also wants the Federal Communication Commission (FCC) to assess whether the State AI’s regulations interfere with the ability to carry out their responsibilities and authorities under the 1934 Communications Act. If Trump somehow enforced this part of the project, it would turn the FCC into something very different from what is today.
“The idea that the FCC has authority over artificial intelligence,” said Codi Wenzke, senior policy lawyer of the American Civil Liberties Union. “Traditionally, it has no jurisdiction over things like website or social media. This is not a privacy agency, and thus, looking at the fact that the FCC is not a complete service technology regulator, it is really difficult to see how it has the authority over AI. “
Hall notes that this part of Trump’s plan is especially worrying about how the president has restricted the agency’s independence. In March, Trump illegally fired two FCC Democratic Commissioners. In July, Anna Gomez, the only rest of the commission’s Democrat, accused the Republican chair Brandon Car of “agency” to “surrender” to silence critics to the “agency”.
“It is surprising that the president is choosing to go alone and is unilaterally trying to impose a backdoor estate Mauritime through the FCC, and is looking for federal funds and recognizes his own laws that may be related to AI,” said Venezke.
AGUR BONFIC for Angate
On Wednesday, the president also signed three executive orders to start his AI agenda. One of them, titled “Awakening in the Federal Government”, limits federal agencies to just acquiring AI systems that are “to find the truth” and free from theory. The order states that, “LLM will be neutral, non -party tools that do not manipulate in response to ideological dagas like DEI.” “LLM will prefer historical accuracy, scientific inquiry, and purpose, and recognize uncertainty where reliable information is incomplete or inconsistent.”
The disadvantages of such a policy should be clear. “The plan to determine what is absolute truth and ideological neutrality is a despair,” Venezke said. “Obviously you don’t want government services to be done politics, but the mandate and executive order is not viable and quit serious questions.”
Hall added, “It is very clear that their purpose is not neutral.” “In fact, ideological prejudice is needed, what they have, and what they call neutral. In keeping with it, what is really needed is that the LLM obtained by the federal government includes their own ideological prejudice and slate.”
Trump’s executive order produces a discretionary political test that companies like Open AI should pose a risk to lose or risk government contracts – AI is actively courting firms. Earlier in the year, the Open launched the Chat GPT Government, designed to use a version of its chatboat for the use of a government agency. Zee announced Grook for the government last week. “If you are creating LLMs to meet the government’s purchase requirements, it is a real concern that it is about to reach a wider private use,” said Venezke.
If the Trump administration should find a way to empower the FCC to manage the AI somehow, then according to these reactionary parameters, AI products facing consumers are more likely. Under the Brandon Car, the Commission has already used its regulatory power for strong companies so that they are in line with the President’s position on diversity, equality and participation. In May, Verizon received the FCC approval for his $ 20 billion integration with the Frontier after promising to eliminate all DEI methods. The Skydance has also made a similar pledge to stop the approach of Paramount Global’s $ 8 billion.
Even without direct government pressure to do so, Elon Musk’s Grook Chatboat has shown twice this year, how the “maximum truth -seeking” result looks like. First, in mid -May, he made unprecedented claims about “white genocide” in South Africa. Recently, it became a complete “Machaterler” and he turned towards the Jewish enmity.
According to Venezke, the entire plan is probably “illegal” to prevent Trump’s states from organizing the AI, but this is a little comfort when the president has actively provoke the law many times to count in less than a year in his second term, and the courts have not always decided against their behavior.
“It is possible that the administration will easily read the directives of the AI Action Plan and will move forward thinking about the FCC’s jurisdiction, when federal programs have in fact created a dispute with state laws, and this is a very different conversation.
