As artificial intelligence development races on toward the unknown, the known for AI—whether it be rampant leftist political bias, hallucinations or outright censorship—continues to show a concerning trend of negative externalities for users.
But the reasons to run, not walk, toward artificial general intelligence (AGI) and artificial superintelligence (ASI) are many: massive economic, strategic and societal advantages; national security risks and global dominance concerns; and overall efficiencies in scientific progress, among others.
Thus, global tech firms are continuing to race. As they do, there are important policy pursuits that should be avoided going forward. Below is a non-exhaustive list of recent flashpoints in tech development policy that could rear their heads again in the not-so-distant future:
Moratoriums on State AI legislation
-
The U.S. House passed a 10-Year moratorium on AI legislation by states. One key argument for this policy was to make sure that America does not lag behind other nations in the AI arms race, keeping U.S. adversaries like China in check. But what the 10-year moratorium didn’t do was allow states to protect their states’ citizens’ free speech liberties, like Texas and Florida have tried to do by passing anti-censorship laws in recent years. Instead, Big Tech would be empowered to continue the same sort of censorship it has wrought on Americans in the recent past with near impunity. If the COVID-19 era has taught conservatives anything, it should be that states must not be shackled by federal regulations or legislation that prevents them from protecting their constituents' freedoms.
-
Senate Republicans attempted to pass a compromise on the 10-year moratorium via a five-year one-way ratchet amendment. For varying reasons, some seeking to protect children and others concerned with states’ rights and federalism, opposition to the 10-year moratorium emerged and a far worse, five-year one-way ratchet amendment was proposed in the waning hours before passage of the One Big Beautiful Bill Act.
Under the proposed amendment, state statutes that prohibited censorship by AI companies — statutes the U.S. Supreme Court recently upheld — would be all but nullified. In the same fell swoop, state regulations that incentivize censorship, such as legislation pushed by leftist politicians in the state of New York or elsewhere, would remain. Thankfully, and at nearly the last possible moment, the five-year one-way ratchet amendment was also put to bed, albeit seemingly not for pro-free speech purposes.
But don’t be fooled, as Politico noted on July 2, the AI moratorium is likely to make its way back onto the scene. While it is possible that a provision exists that could thread the needle for the U.S. to outpace China while preserving states’ rights to prevent censorship, conservatives must be armed with information and ready to defend liberty in the event that legislation like the most recent iteration rears its head again.
Censorship by Another Name
-
Beware of efforts by well-intentioned people pushing “safety” and “age verification” laws, like the Senate Democrat version of the Kids Online Safety Act (KOSA). There are good ways and bad ways to provide an online environment that is safe for everyone, including for America’s kids, but using “safety” as a cudgel while leaving openings for censorship is not the way to go about it. The House version of KOSA does a much better job at striking such a balance. Instead of providing censorship loopholes, the House version stripped the controversial “duty of care” provisions pushed by Democrats in the Senate version of the bill.
Similar to the Senate version of KOSA, some age verification laws across the country have a tendency to provide greater incentives for Big Tech to censor. As MRC President Emeritus Brent Bozell said about online censorship and child safety: “Protecting children is extremely important, but we should not fall into the trap of allowing Big Tech to censor us, which we know will happen on issues like gender ideology and abortion.”
-
X is adding AI to its (less-bad, but still not good) Community Notes censorship feature. In an X post last Tuesday, X CEO Linda Yaccarino bragged that the X platform would be “advancing the frontier again,” ushering in “‘AI Note Writers.’” As Fox News columnist David Marcus put it for MRC Free Speech America, Community Notes are “a crowdsourced alternative to professional fact-checking, and while some conservatives appear to like the results better, the warning labels are still a form of censorship, albeit by a different name.” Adding AI to the crowdsourced so-called fact checks on X, AI-created notes will be added to the mix if enough users rate them as “helpful” to the conversation, “accelerat[ing] the speed and scale of Community Notes.”
If the free speech concerns with Community Notes were not immediately apparent, this new AI acceleration coupled with X owner Elon Musk’s and Yaccarino’s penchant for parroting former Twitter co-founder Jack Dorsey’s anti-free speech “Freedom of Speech, Not Reach” philosophy should make it more clear. As the saying goes: If a tree falls in the woods and no one is there to hear it, does it make a sound?
Avoiding Big Tech Antitrust Exemptions
-
Avoiding another iteration of the media cartel bill — this time with AI. Not so long ago, politicians on the left argued that in order to preserve free speech we had to squash it. In doing so, they pushed the Journalism Competition and Preservation Act of 2022. This “media cartel bill” would have created “a four-year safe harbor from antitrust laws for print, broadcast and digital news companies to allow news outlets to jointly negotiate payment terms for content distribution on social media,” as reported by MRC Free Speech America. Laudable as it was for the left to argue for an end to Big Tech cartels, this was a horrible attempt. As former MRC VP Dan Gainor simply put it: “Cartels are bad. Trying to fix one set of cartels by creating another set of cartels is a bad solution.”
And as if almost on queue, here they go again. The left, instead of outright calling for cartels, AI companies and legacy media appear to be recreating the same cartels — but this time by using AI.
Using copyright as a blunt instrument to tie the hands of AI companies, the flailing legacy media are now pushing AI companies to enter into so-called licensing agreements. Such arrangements only serve to provide the left control over what Americans see and hear from AI large language models, or chatbots (Big Tech’s new search tool).
Enter Trump AI Czar David Sacks to address the left’s AI copyright gambit.
“[What] AI models do in pre-training is they take millions of texts, millions of documents and they understand the positional relationship of the words and they translate that into math, into a vector space called positional encoding. That is a transformation of the underlying work. In the same way that a human can read a body of works and then come up with their own point of view, that is basically what AI models are doing.”.
Sacks went on to explain that AI companies need to be allowed to use as many sources as possible in training to be as robust and helpful as possible. Sacks said, “So, if AI models violate someone's copyright by outputting something that's identical, then, obviously, that's a violation, but if all they're doing is transforming the work, they're doing positional encoding and then coming up with their own unique work product. This judge [Alsup] said that that is not a violation of copyright.”
And while Sacks argued thoughtfully about how placing too many restrictions on AI learning would weaken AI models, potentially making American companies fall behind China, he did not mention the disturbing Big Tech-legacy media trend toward “exclusivity contracts.”
Some AI companies, most notably OpenAI, may be violating antitrust laws by signing what amount to exclusivity contracts with left-leaning publications like The Associated Press, The Washington Post, Axios, Time, Vox, The Atlantic and The Financial Times, among others.
The problem with such exclusivity contracts is plain: AI companies training their models on legacy media content, particularly if combined with restrictions preventing those same companies from training on other sources, means that AI models will provide biased, left-leaning responses (and citing to the legacy media sources on which they were trained).
Conservatives may disagree on how to rein in Big Tech while beating China to the punch on AI, but all can agree that safeguarding liberty, especially America’s first freedoms, must be top of mind in every policy pursuit.
Conservatives are under attack! Contact your representatives and demand that Big Tech be held to account to mirror the First Amendment while providing transparency, clarity on hate speech and equal footing for conservatives. If you have been censored, contact us using CensorTrack’s contact form, and help us hold Big Tech accountable.
