Despite the rallying of part of the Silicon Valley, the Trump administration does not intend to give a blank to the Big Tech, even if its position on the AI remains ambiguous.
Are generative-based chatbots, like Chatgpt, constitute a danger to the mental health of minors who use them? This is what the Federal Trade Commission (FTC), the American authority responsible for the application of consumer and competitive practices, to determine.
For this, the agency has asked the tech giants who operate such chatbots, including OpenAi, Meta or Character. To send it internal documents which will allow it to assess the risks and plan future actions accordingly.
Suicide of several young users
The American authorities are under pressure from parental organizations and minor defense groups who require better supervision of these chatbots to protect the youngest. These pressures follow in particular to several media affairs involving young people who have committed suicide after having established very strong friendly, even romantic relationships, with their chatbot. Last year, a 14 -year -old young American man died after he entered an obsessive relationship with a character chatbot. Ai modeled on the personality of Daenerys Targaryen, a character from the Game of Thrones television series. Character.a allows the user to design chatbots based on a fictitious or real personality. The young man’s family brought the start-up a trial, believing that the chatbot amplified the depression of it and pushed it to suicide.
Several American policies, right and left, have also seized the subject. Rebecca Bauer-Kahan, an elected Democrat of California, asked at the end of August a closer supervision of the practices of the companies of AI after the suicide in the Golden State of a 16-year-old who became addicted to her chatbot. Josh Hawley, senator of the Missouri embodying the populist wing of the Republican party, and Marsha Blackburn, elected republican of Tennessee, for their part launched an investigation against Meta after internal documents of the company revealed in August that she authorized her chatbots to have romantic conversations with children.
Meta responded to criticisms by denying any erotic conversations between her chatbot and minor users, and by setting up a battery of new rules prohibiting AI from approaching a certain number of subjects with them, including suicide, self -control and food disorders. Openai has also announced working on his models to better teach them to recognize the signs of mental and emotional fragility in its users, and to respond to them adequately.
The risks posed by AI are also a cause in which the First Lady was involved, Melania Trump, who at the beginning of the year helped the passage of a law against hypertructures.
In a very conflictual American political landscape, the protection of minors against the drifts of Big Tech is one of the few subjects that transcends cleavages in politics. In July, the Senate thus voted with a large majority the Kids Online Safety Act, a law aimed at better protecting young people on social networks, promoted by Marsha Blackburn and the Democratic Senator of Connecticut Richard Blumenthal. The law must now be passed by the House of Representatives to complete its legislative career and be signed by the President.
Trump against AI “Woke”
If it transcends political cleavages, the desire to better supervise AI is however not shared by the entire Trump administration. This also includes a libertarian branch, close to Silicon Valley, which sees attempts to impose safeguards on technology the risk of slowing down its progress, for the benefit of geopolitical rivals of the United States like China, which progresses by giant in this area. This speech affects a sensitive cord with the president and has already given rise to several deregulations favorable to IA companies.
At the start of his second mandate, the president thus canceled a decree signed by Joe Biden who fixed certain security standards to be observed for the major AI models and provided for the drafting of a report evaluating the risks posed by AI on American users and the labor market. In a highly symbolic way, the Trump administration also renamed AI Safety Institute, launched in 2023 by Joe Biden to supervise the way the public authorities can and must use AI. It is now called the Center for AI Standards and Innovation, and its mission has been refocused on the imperative “to ensure American domination over AI”. Donald Trump has also signed a decree aimed at promoting the use of AI in the educational sphere.
The president, who must endeavor to satisfy the technophile branch as well as the populist wing of his clan, does not intend to abandon the regulation of the AI, as shown in the investigation opened by the FTC. On the other hand, he seeks to redirect this regulation in order to make consensus within his own camp. Hence a focus on the protection of minors, a relatively consensual theme.
Conversely, the wishes to impose an AI without racist or sexist bias via regulation are vigorously rejected. Last June, Donald Trump signed a decree aiming, according to his own words, to end “the Woke”. This decree requires in particular IA companies that receive public money that they conceive of neutral political models and “devoid of ideological dogmas such as diversity, equity and inclusion”.
In practice, this decree concerns most American Giants of AI: Openai, Anthropic, Google, Meta and Xai have all won contracts from the Pentagon, as part of which they affect public money. The policy carried out by the Trump 2.0 administration on AI is thus the reflection of the different trends, sometimes very different, which compose it.
