When Google’s AI chatbot, Gemini, produced images of a black George Washington and declared that labeling “all individuals with pedophilic interest as ‘evil’ is inaccurate and harmful,” it exposed a deeply disturbing, “intentionally” designed bias at the big tech behemoth.
Now, House Republicans, led by Ohio Rep. Jim Jordan, want Google to disclose the extent to which the development of Gemini was “influenced” by “the Biden White House.”
In a Saturday letter addressed to Washington-based King & Spalding law firm attorney Daniel F. Donovan, Jordan wrote, “In light of new reporting regarding how Alphabet intentionally biased its AI model, Gemini, by giving it instructions that distort the results shared with Americans who use the model as a source for information, we write to inform you that the Committee views the subpoena as covering material relating to this technology.”
#BREAKING: @Jim_Jordan Demands Information on the Biden Administration’s Role in Google’s “Woke” Gemini AI
The Committee is investigating how and to what extent the Executive Branch has colluded with Big Tech and other intermediaries to censor Americans’ speech. pic.twitter.com/R2GKSpsRnG
— House Judiciary GOP (@JudiciaryGOP) March 2, 2024
For context, “King & Spalding’s highly regarded Congressional Investigations lawyers have a level of experience and effectiveness that is unmatched in representing and helping protect the interests of clients as they navigate the unique aspects of an investigation by Congress,” according to the international firm’s website.
One of those clients is Alphabet, the parent company of Google.
“Recent reporting alleges that the Executive Branch, most notably the Biden White House, may have influenced the development of Alphabet’s Gemini AI model,” Jordan told Donovan.
The House Judiciary Committee chair pointed to an Executive Order issued by Biden in which he proclaimed, “Artificial Intelligence policies must be consistent with my Administration’s dedication to advancing equity and civil rights.”
“Artificial Intelligence systems deployed irresponsibly have reproduced and intensified existing inequities, caused new types of harmful discrimination, and exacerbated online and physical harms,” President Biden stated in the October 2023 order. “My Administration will build on the important steps that have already been taken — such as issuing the Blueprint for an AI Bill of Rights, the AI Risk Management Framework, and Executive Order 14091 of February 16, 2023 (Further Advancing Racial Equity and Support for Underserved Communities Through the Federal Government) — in seeking to ensure that AI complies with all Federal laws and to promote robust technical evaluations, careful oversight, engagement with affected communities, and rigorous regulation.”
My Executive Order on AI is a testament to what we stand for:
Safety, security, and trust. pic.twitter.com/rmBUQoheKp
— President Biden (@POTUS) October 31, 2023
The “Safe, Secure, and Trustworthy Development and Use of Artificial Intelligence” Executive Order “states that its guidelines will purportedly ‘advance[] equity,'” Jordan wrote. “A recent report by the company’s Gemini Team stated that its new AI tool was evaluated by external groups, which were selected based on their alignment with the Biden White House Executive Order and ‘White House Commitments.’ Alphabet claimed that it worked with these external groups ‘to help identify areas for improvement,’ including ‘societal risk,’ such as alleged ‘representational harms.'”
The request for information, Jordan told Donovan, is a matter of protecting free speech.
“Congress has an important interest in protecting and advancing fundamental free speech principles,” he wrote. “To develop effective legislation, such as the possible enactment of new statutory limits on the Executive Branch’s ability to work with Big Tech to restrict the circulation of content and deplatform users, the Committee must first understand how and to what extent the Executive Branch coerced and colluded with companies and other intermediaries to censor speech.”
Gemini AI ‘fiasco’ sparks renewed concerns of deep-seated political bias at Google https://t.co/8JqfxQmrQV via @BIZPACReview
— BPR based (@DumpstrFireNews) March 1, 2024
“Given that Alphabet has censored First Amendment-protected speech as a result of government agencies’ requests and demands in the past, the Committee is concerned about potential First Amendment violations that have occurred with respect to Alphabet’s Gemini model,” he wrote.
“The purpose of this letter is to inform you that communications and documentation pertaining to the Executive Branch and other third-parties relating to the creation, training, and deployment of the Gemini AI models are within the scope of the Committee’s subpoena,” Jordan told Donovan. “The subpoena requires Alphabet to produce, among other things, material concerning Alphabet’s engagement with the Executive Branch and Alphabet’s decisions and policies regarding content moderation.”
“Please produce all documents and information as soon as possible,” Jordan requested, “but no later than 5:00 p.m. on March 16, 2024.”
Comment
We have no tolerance for comments containing violence, racism, profanity, vulgarity, doxing, or discourteous behavior. If a comment is spam, instead of replying to it please click the ∨ icon below and to the right of that comment. Thank you for partnering with us to maintain fruitful conversation.
