Leaving the career for artificial intelligence. This is the goal of Donald Trump, who presented his action plan at AI on July 23. A three -axis -based plan, whose hunt for the US president calls “ideological biases”, mainly initiatives that promote the representation and inclusion of minorities (DEI).
The US President attacks isolated to isolated too “awakened”, wishing to guarantee freedom of expression in AI models. For the United States to be a leader in the AI field, Donald Trump also considers necessary to eliminate regulations that “unnecessarily hinder the development or deployment of AI.”
A plan that may not please the European Union (EU), which prepares its Act Law and is also in full negotiation with the United States over customs duties.
Freedom of expression
For Donald Trump, chatbots should allow users to express their opinions freely while they are shown. “Americans will need reliable results, but the integration of ideological biases or social programs in AI models can distort the quality and precision of the results,” said the White House in its action plan.
As part of diversity, equity and inclusion initiatives, conversational agents such as Chatgpt can manipulate racial or sexual representation in their results or even integrate concepts such as changes in sex, she reproaches.
To prevent this from happening again, Donald Trump will force the government to sign contracts only with developers of large language models that show objectivity.
As indicated, this is equivalent to having models similar to Grok, the AI Elon Musk that has fewer limits than others and has made controversy on several occasions, especially with anti -Semitic comments.
Suffocating regulations
Without a doubt, this change will be criticized by the European Union, which he said in early July “to take very seriously” the controversy about Grok and be in contact with the social network X after the chatbot has chained the anti -Semitic, racist and conspiracy comments. The European Commission also said that “would examine” a letter received from Poland and the appeal to open an investigation into Grok after this case.
To this is added the law, the European law that will impose rules to the AI models for general use (Chatgpt, Gemini …). In particular, they will have to “evaluate and relieve potential systemic risks.” Regulations that are in Donald Trump’s viewfinder. In his action plan, the US president affirms Black on White that “to maintain his world leadership in AI, the American private sector must be released from bureaucratic heaviness.”
The president believes that “AI is too important to be suffocated by the bureaucracy at this early stage.” This is the reason why one of the measures of your action plan will consist of a collaboration with all federal agencies to “identify, review or repeal regulations, rules, memoranda, administrative orders and inter -infinitions that unnecessarily hinder the development or deployment of AI”.
Confrontation with the EU
It is no coincidence that the White House presented this plan on July 22. Because the United States and the European Union have until August 1 to find an agreement on customs duties. Donald Trump could impose high customs taxes in retaliation if Brussels does not give up.
Two complex scenarios. Because, on the one hand, today’s chatbots are mostly Americans. Even Mistral, which is French, wants to extend to the United States. In June 2024, he raised 600 million euros of mainly American investors such as NVIDIA or IBM, indicating that this sum will be used to “expand its presence internationally”, especially in the United States after having opened a branch in California.
On the other hand, Europe cannot miss the AI. “Imagine this same discussion in the 2000s, where we say that we will prohibit the Internet in the territory. Imagine in what state we would be today if we had not had the Internet for 20 years compared to the rest of the world,” said the lawyer, who considers it necessary for the EU to illuminate its regulation over AI, which is currently a great brake “, even for French companies.
Opportunistic companies
Having taken advantage of the arrival of Donald Trump in power, American companies did not wait for the White House Action Plan to establish changes. In February, Openai modified its model specification, a document that governs the way in which its AI models behave, to “the search for truth together.” An objective that includes the fact of the latter of “not having an agenda” and allowing users to express themselves freely.
For its part, Meta recently refused to sign the “AI AI” Code. A voluntary code that, according to the Commission, demonstrates that IA models suppliers for general use respect the obligations of the law.
“Europe is committed to the wrong route in terms of AI. We have carefully examined the Code of the European Commission of Good Practices for AI models for general use (GPAI) and goal will not sign it. This code introduces a certain number of legal uncertainty for models developers, as well as measures that far exceed the scope of Law AI,” said Joel Kaplan, linked.
It remains to be seen if the EU yields to Donald Trump’s pressures.
Source: BFM TV
