Under Trump, AI Scientists Are Told to Remove ‘Ideological Bias’ From Powerful Models

The National Institute of Standards and Technology (NIST) has issued new instructions to scientists that are parted with the US Artificial Intelligence Protection Institute (AISI) that “AI Protection,” “Responsible AI,” and “AI Fairness” expected to reduce humanitarian requests to reduce humanitarian requests. By introducing a request for.

This information comes as part of an updated cooperative research and development agreement for members of the AI ​​Protection Institute sent in early March. Previously, this agreement encouraged researchers to contribute to the technical work that could help identify and resolve discriminatory models related to gender, race, age or wealth discrimination. These national biases are very important because they can directly damage the end users and minorities and economically disadvantaged groups.

The new agreement removes the developer tools “for content authentication of” synthetic content labeling “as well as” synthetic content labeling “to track the content of the content and its introduction. It emphasizes the first to keep the United States on the first, to develop a working group to develop testing tools “to expand the Global AI position in America”. “

A researcher at an organization working with the AI ​​Protection Institute said, “The Trump administration has removed security, fair, wrong information and liability,” a researcher working with the AI ​​Protection Institute, who did not want to be named for revenge, said.

Researchers believe that ignoring these issues can damage regular users, perhaps not to verify the discriminatory algorithm based on income or other population. “If you are not a technical billionaire, it will lead to the worse future for you and you who are careful. AI hope that wrong, discriminatory, insecure and irresponsible will be deployed, ”the researcher claimed.

Another researcher who has worked with the AI ​​Protection Institute in the past says, “It’s wild,” “What does even mean human development mean?”

Elon muskWho is currently leading A controversial attempt The President has criticized OpenAI and AI models created by Google to minimize government expenditure and bureaucracy for Trump. Last February, he posted a meme on X so that Gemini and Opena were identified as “racist” and “awake”. He is often Quoted An event where Google’s model debuted that it was argued about whether it would be wrong to make someone a mistake even if it could stop any nuclear apocalypse – this is a very impossible situation. As well as Tesla And SpaceXJai operates an AI company that competes directly with OpenAI and Google. A researcher who advised Jai recently recently created a fancy technique to change the political tendency of big language models, eg Report By wired.

A growing agency of the study shows that political bias in AI models can affect both Liberal And conservative. For example, A study of algorithms to recommend Twitter Published in 2021, it has shown that users were more likely to show the right-risk view of the platform.

From January, the musk is so -called Government skill department (Doses) The US government is effectively moving, effectively firing civil servants, breaking the expenses and creating an atmosphere that is considered a hostile environment with the Trump administration’s goals. Some government departments, such as the DEI mentioned in the Department of Education, have archived and deleted the documents mentioned by the DEI. Doz has also targeted NIST, the main AISI’s main company in recent weeks. Dozens of employees have been dismissed.