Political trends and circumstances change, as do US administrations – but the House Judiciary Committee chaired by Jim Jordan continues to “soldier on” in its multi-year, comprehensive bid to get to the bottom of the activities by the Biden-Harris White House aimed at pressuring tech companies to its political advantage.
In the past, these investigations produced some spectacular results – such as Meta CEO Mark Zuckerberg publicly admitting to his company relenting to that pressure, stating he regretted that – and that the tech giant would reverse the policies that facilitated compliance with the former government.
The latest set of the Committee’s subpoenas concern companies developing AI tech. The subpoenas have been sent to Adobe, Alphabet, Amazon, Anthropic PBC, Apple, Cohere, International Business Machines Corp., Inflection AI, Meta, Microsoft, Nvidia, Open AI, Palantir Technologies, Salesforce, Scale AI, and Stability AI.
We obtained a copy of one of the letters for you here.
The Committee wants all documents and communications that the previous administration had with these companies concerning content “moderation and suppression” – i.e., collusion with the aim of censoring lawful speech – to be preserved and presented. The timeframe is January 2020 to January 2025.
The Committee’s looking into possible government pressure to use various forms of available AI to carry out censorship, either by the firms that received the subpoenas, or any third-parties involved.
While announcing the subpoenas, the Committee noted its work, during the term of the previous Congress, to shed light on the way the former US government put pressure on companies behind social networks to censor lawful speech, which touched on the use of AI, that is, algorithms programmed to suppress speech.
Such implementations were declaratively implemented to deal with “harmful bias” and “algorithmic discrimination” as AI models get developed – but now the Committee wants a much closer look into how this may have aided the Biden-Harris administration’s overall censorship effort.
The goal of the Committee’s work here is to “inform legislative reforms to protect Americans’ civil liberties in light of developing AI technologies,” a press release said.
The October 2024 memorandum to the National Institute of Standards and Technology (NIST) is one of the instances singled out in the subpoenas – the memorandum’s goal being “to establish an enduring capability to lead voluntary unclassified pre-deployment safety testing of frontier AI models on behalf of the United States Government.”