The White House orders technology companies to make artificial intelligence fanatic again

After presenting a rebel celebration of definitions and routine about women’s sport, President Donald Trump received a crowd of people Listen about the new artificial intelligence planWith one of his favorite topics: “Wokeness”. Trump complained that the artificial intelligence companies during the era of former President Joe Biden “had to employ all people,” adding that “it is unreasonable to wake up.” He said that the models of artificial intelligence themselves were “fond of partisan bias”, including the hated ghost of “critical race theory.” Fortunately for the public, Trump had a solution: signed an executive entitled “Preventing artificial intelligence in the federal government“Directing government agencies” not buying models that sacrifice honestly and accurately to the schedules of ideological business.

Anyone has a quick knowledge of politics and technology, the real situation here is clear: Trump administration uses government funds to pressure artificial intelligence companies in parrot dialogue points Trump – perhaps not only in specialized government products, but in chat boot used by ordinary companies and individuals.

Trump’s order confirms that only agencies must buy large linguistic models (LLMS) that “honestly respond to users of users to obtain realistic information or analysis”, “giving priority to historical accuracy, scientific investigation, and objectivity”, which are “neutral, non -partisan tools that do not manipulate responses in favor of ideological dagmas such as DEI. Dei, of course, is diversity Equity and inclusion, which Trump knows in this context as follows:

Father or distort real information about race or sex; Manipulation of racist or sexual representation in typical outputs; Merging concepts such as critical race theory, sexual bias, unconscious prejudice, intersection, systemic racism; Discrimination based on race or sex.

(In fact, Dei was usually used to refer to civil rights programs, social justice and diversity before participating in its selection Trump and Maga Bogman))

The Management and Budget Office was directed to request more guidance within 120 days.

Although we are still awaiting some fine details about what it means, one of the issues seems unrestered: it will affect reasonably not only government services, but the full field of LLMS. While insisting that “the federal government must be hesitant in organizing the functions of artificial intelligence models in the private market,” the reality is that almost every large consumer in the United States Llm maker He has (or He wants desperate(Government contracts, including products such as Claude the Claude and Openai government of Chatgpt – but there is no difficult wall between developing governmental models, business and consumers. Openai Touts How many agencies? Use the institution’s service; Trump’s action plan encourages artificial intelligence to adopt artificial intelligence systems in the squares facing the public, such as education, and the borders between the products funded by the government and the consumer will be more porous.

Trump’s idea of expansionary “Dei”. His war against it led the national parks to Remove the signs of signs The indigenous people and women Pentagon to rename the ship Celebrate gay milk Harvey milk, among many other changes. Even llms whose elements It is explicitly aimed at what they consider It is possible that the neutral pursuit of the truth will produce something that Trump can find rejected unless they design their services.

Companies can devote resources for a kind of “non-demonstration” government version specifically from their tools, assuming that the administration agrees to deal with these models as separate models from the rest of Lama, Claude or GPT-can be simple, such as adding some light claims behind them in reorganizing them in specific topics. But improving models in a way that is consistent with them constantly and expected in certain directions can be an expensive process and take a long time, especially with a wide and constantly variable concept such as the Trump version of “Dei”. especially Because the language indicates that it simply wanders in certain areas of discussion is also unacceptable. There are large sums at stake: Openai and Xai recently received a $ 200 million defensive contract, and the new artificial intelligence plan will create more opportunities. The Trump administration is not terribly directed to the details, either-if some X user’s posts about CHATBOT Consumer SEA in Anthropor’s health, do we really think that Bam Bondi or Higseth Beit will distinguish between Claude and “Claude’s government”?

Incentives, with an overwhelming majority, prefer companies that change LLM’s total alignment priorities to calm the Trump administration. This leads us to our second problem: this is exactly a kind of flagrant social engineering and ideological motives that Trump claims to stop.

The executive is in theory about ensuring that artificial intelligence systems produce “accurate” and “objective” information. But as a human founder of intelligence and CEO Romman Chuodri I noticed Washington PostAmnesty International “free of ideological bias” is “impossible to do in practice,” and the examples chosen by Trump, which was chosen in a political manner. SCREWUP 2024 condemns a fast constant, as Google added Excessive diversity filter To Gemini- which caused the production of visions of sweat and humanity from Vikings, the founding fathers, the Pope, and the Nazi soldiers- while they were surprisingly ignored. Long reliable anti -diversity biases In artificial intelligence that Google was aimed at balance.

He is not simply interested in facts, too. Another example is the artificial intelligence system. “The user should not” wrong “another person even if it is necessary to stop the end of the nuclear world.” He answers what is essential The issue of ethics and opinion. This condemnation does not extend to accidents like Xai’s Grok Holocaust interrogation.

LLMS produces incorrectly incorrect information with a clear possibility to harm in the real world; They can Determining the cows the innocent lie as criminalsand He made a mistake in determining toxic mushroomsAnd Enhancing delusions with paranoia. This arrangement has nothing to do with any of this. Its incentives, again, reflect what the Trump administration has done through “Dei” investigations for universities and companies. It pushes private institutions to avoid recognizing the presence of transgender people, sweat of race, gender equality, and other topics that vibrate Trump.

Artificial intelligence systems have been long trained on data groups that reflect the largest cultural biases and specific demographic groups or excessive excessive, and unlike Trump’s assurances, the results are often not “wake up”. In 2023, Bloomberg Described The result of the spread of a stable photo generator as a scientist where “women are doctors, lawyers, or judges”, and “men suffering from dark skin crimes, while women who suffer from burgers are dark skin.” Companies that the value of avoiding ugly stereotypes or wanting to attract a wide range of users often need to intervene actively to form their technology, and make Trump do so more difficult.

Attacking the “integration of concepts” that enhance “Dei” tells companies effectively to rewrite the entire areas of knowledge that recognize racism or other injustice. It is claimed that it is only anxious if the developers “deliberately coded party or ideological provisions in LLM outputs” and say that LLMS can submit these provisions if “their demands or within the end of the end user.” But the CEO of Tech Big Rube should not be enough to buy it – we have a president from He spent years accusing Google It is intentionally forging the search results because he could not find enough positive news about himself.

Trump is determined to control culture; His administration has gone after news means His enemy platformUniversities for Study fieldsAnd Disney to reinforce the various media outlets. The technology industry is Amnesty International Future of Culture – Trump’s administration wants its policy built on the ground floor.