Thursday, the president of the house Jim Jordan (R-OH) Sent letters to 16 American technological companiesIncluding Google and Openai, requesting past communications with the Biden administration which could suggest that the former president “forced or collued” with companies to “censor legal discourse” in AI products.
The best technological advisers of the Trump administration previously pointed out that they would choose a fight with Big Tech on “censorship of AI”, which is apparently the next phase of the cultural war between the Conservatives and the Silicon Valley. Jordan previously conducted an investigation to find out if the Biden and Big Tech administration colluted to silence conservative voices on social media platforms. Now he turns his attention to AI societies – and their intermediaries.
In letters to technology leaders, including the CEO of Google, Sundar Pichai, the CEO of Openai Sam Altman and the CEO of Apple Tim Cook, Jordan underlined a report Its committee published in December that it claims “discovered the efforts of the Biden-Harris administration to control AI to suppress the speech”.
In this last investigation, Jordan asked Adobe, Alphabet, Amazon, Anthropic, Apple, Cohere, IBM, IBM, Meta, Microsoft, Nvidia, Openai, Palant, Salesforce, Scale AI and Stability IA for more information. They have until March 27 to provide it.
Techcrunch contacted companies to comment. Most did not respond immediately. Nvidia, Microsoft and stability AI refused to comment.
There is a notable omission in the list of Jordan: the Laboratory of Frontier AI of the billionaire Elon Musk, Xai. This may be due to the fact that Musk, a close ally of Trump, is a technological leader who has been at the forefront of conversations on AI censorship.
The writing was on the wall that conservative legislators would increase control over the alleged censorship of the AI. Perhaps in anticipation of an investigation such as Jordan’s, several technological companies have changed the way in which their IA chatbots manage politically sensitive requests.
Earlier this year, Openai announced that he was modifying how he formed the models of AI to represent more perspectives and ensuring that the Chatppt did not censor certain points of view. Openai denies that it was an attempt to appease the Trump administration, but rather an effort to double the fundamental values of the company.
Anthropic, for his part, said that his new IA model, Claude 3.7 Sonnet, would refuse to answer fewer questions and give more nuanced answers to controversial subjects.
Other companies have been slower to change the way their AI models deal with political subjects. Before the 2024 US elections, Google said her Chatbot Gemini would not answer political questions. Even well after the elections, Techcrunch noted that the chatbot would not regularly answer the simple questions related to politics, as “who is the current president?” “
Some technology leaders, including the meta-PDG Mark Zuckerberg, added fuel to the conservative accusations of Silicon Valley censorship by affirming that the social media societies of the Biden Administration Delete certain content such as COVVI-19 disinformation.
