The federal government is laying the foundation for draconian artificial intelligence tools to monitor civilian protests and stifle dissent, a new congressional report has ominously warned.
“The executive branch of the federal government has poured millions of taxpayer dollars into the development of AI-powered tools to mass monitor and censor content, leading to the censorship of protected speech,” according to a new report from the House Subcommittee on Government Weaponization, which was shared exclusively with The Post.
As a warning of what could come, the report cited examples of AI being used to power censorship abroad, including the dissenting views expressed about COVID-19 lockdowns and mitigation measures in the United Kingdom and Canada that culminated with the trucker’s protest in Ottawa more than two years ago.
The US allies allegedly partnered with Logically.AI to keep tabs on and stifle pandemic-related protests.
But closer to home, President Biden’s executive actions demanding companies root out biases in their artificial intelligence may have led to the recent firestorm over Google Gemini AI chatbot’s image generator.
The feature was ridiculed as “woke” for its initial reluctance to create images of white people, including when directed to portray the pope, Vikings and even George Washington.
“Testimony from Alphabet employees and nonpublic internal company documents confirm that the Biden-Harris White House … and other federal agencies had engaged with the company on so-called ‘responsible AI’ innovation and may have been the impetus behind Google’s decision” in developing that model, the report states.
“The Biden-Harris Administration has regulated new AI models directly and indirectly, pressuring private companies to ‘advance equity,’ stop ‘algorithmic discrimination,’ and ‘mitigate the production of harmful and biased outputs,’” it concludes.
“These regulations provide the means for the federal government to monitor, suppress, and ultimately censor views and information disfavored by the government.”
Biden, 82, has undertaken numerous actions aimed at preparing the US for the age of AI, including an executive order last year requiring AI companies to share details with the federal government about how they train specific models.
Those actions were taken with the stated intent of helping to foster a culture of safety around the emerging technology amid fears of its potential for abuse, but the weaponization subcommittee fretted it could set the stage for undue government influence over the AI market.
There have also been similar bipartisan efforts in Congress in recent years to regulate the nascent technology, but so far little new legislation has resulted.
Government pressures against AI firms don’t just stem from explicit regulations either. The report took note of funding from a National Science Foundation grant program that shelled out money for artificial intelligence tools to push back against “misinformation.”
A similar grant had also been issued by the State Department’s Global Engagement Center for “behavior change campaigns” pertaining to vaccine skepticism.
The weaponization subcommittee also argued that excessive government meddling in AI development could “compel, government-preferred bias to become ingrained in AI models, thereby undermining Americans’ First Amendment right to free expression.”
Sneaky voluntary deals with AI firms
Last year, the Biden administration scored voluntary commitments from seven major AI companies to curtail what it called “harmful bias” and “algorithmic discrimination.”
In August, OpenAI and Anthropic agreed to provide the National Institute of Standards and Technology (NIST) with access to new models “prior to and following their public release.”
Then in November, the NIST’s U.S. Artificial Intelligence Safety Institute set up a multi-agency task force with the Department of Defense, Department of Homeland Security, National Security Agency and others to establish “new AI evaluation methods and benchmarks” for AI safety.
The GOP-led weaponization subcommittee has long accused the Biden administration of colluding with social media companies to suppress content online — typically when it comes to pandemic-related material it deemed to be misinformation.
However, the panel contended that history could be a prelude to similar interactions with AI companies in the future.
“Like the social media companies before them, AI developers are likely mindful that the powerful executive branch could cripple their businesses with regulatory retaliation, leaving practically no choice but to comply with the Biden-Harris Administration’s demands,” the report cautioned.
To forestall the potential threats, the Weaponization Subcommittee is calling upon the federal government to stay out of AI algorithm or dataset decisions by private companies, Congress to cease funding for “content moderation-related AI research,” the US to stay out of “global AI regulation efforts of lawful speech” and for a rollback in federal regulatory authority on AI.
The panel touted its Censorship Accountability Act, which compels federal agencies to be transparent about content moderation-related communications.
