President-elect Donald Trump has surrounded himself with Silicon Valley entrepreneurs, including Elon Musk, Marc Andreessen and David Sachs, who advise him on technology and other issues. are.
When it comes to AI, this team of engineers is pretty much in agreement about the need for rapid development and deployment of AI across the United States. But there is one AI safety issue that this group brings up quite a bit. That is the threat of “censorship” of AI by major corporations. technology.
President Trump's Silicon Valley advisers could turn AI chatbot responses into another battleground for conservatives to fight their ongoing culture war with tech companies.
AI censorship is a term used to describe how tech companies thumb the answers of AI chatbots in order to follow certain politics or push their own politics. Others may call it content moderation, which often refers to the same thing but has very different meanings. Like social media and search algorithms, the goal of AI providing relevant answers to live news events and controversial subjects is constantly changing.
Over the past decade, conservatives have repeatedly criticized Big Tech companies for bowing to government pressure and censoring social media platforms and services. But some technology executives are starting to tone down their public positions. For example, ahead of the 2024 election, Meta CEO Mark Zuckerberg apologized to Congress for bowing to pressure from the Biden administration to aggressively suppress content related to COVID-19. Shortly afterward, Meta's CEO said he had made a “20-year political mistake” by taking too much responsibility for issues outside of his company's control, and said he would not make the same mistake again. said.
But Trump's technology advisers say AI chatbots pose an even bigger threat to free speech and could provide a more powerful tool to influence speech control. Rather than skewing search and feed algorithms toward desired outcomes, like lowering the rankings of vaccine misinformation, technology companies can now provide a single clear answer that doesn't include it.
In recent months, Musk, Andreessen and Sachs have spoken out against AI censorship in podcasts, interviews and social media posts. We don't know how exactly they are advising President Trump, but their publicly stated beliefs could shed light on the conversations happening behind closed doors in Washington, D.C. and Mar-a-Lago. There is sex.
“This is what I believe and what I have been saying to people in Washington: If you think social media censorship is bad; [AI] It could be a thousand times worse,” a16z co-founder Marc Andreessen said in a recent interview with Joe Rogan. “If you want to create the ultimate dystopian world, you would create a world where everything is controlled by an AI that is programmed to lie,” Andreessen said in a separate recent interview with Bari Weiss.
Andreessen also revealed to Weiss that he has spent about half of his time with Trump's team since the election, providing technology and business advice.
“[Andreessen] Former PayPal COO and Craft Ventures co-founder David Sachs said this in a recent post on X, shortly after being named President Trump's AI and crypto czar. “But the timelines diverged and we are now on different paths.”
On All In — host of the popular podcast Sacks along with other influential venture capitalists — President Trump's new AI advisor forces AI chatbots to be politically correct, as show host explains He has repeatedly criticized Google and OpenAI for doing so.
“One of the early concerns about ChatGPT was that it was programmed to launch and not give people real answers about a lot of things. There was censorship built into the answers.” said Sachs on a November 2023 episode of All In.
Despite Sachs' claims, even Elon Musk admits that xAI chatbots are often more politically correct than he would like. That's not because Grok was “programmed to wake up,” but more likely because of the reality of training AI on the open internet. That said, Sachs is making it clearer by the day that “AI veracity” is what he's focused on.
“This is how we can get a black George Washington to join Google.”
The most commonly cited case of AI censorship was when Google Gemini's AI image generator generated multiracial images for queries such as “American Founding Fathers” and “German soldiers from World War II.” However, these were clearly inaccurate.
Image generated by Twitter user Patrick Ganley using Gemini. Image credit: Gemini / Patrick Ganley
But there are other examples where companies have influenced certain outcomes. Just recently, users discovered that ChatGPT would not answer questions about certain names, and OpenAI admitted that at least one of those names triggered its internal privacy tools. At another point, Google and Microsoft's AI chatbots refused to say who won the 2020 US election. During the 2024 election, nearly all AI systems except Perplexity and Grok refused to answer questions about the election results.
In some of these examples, tech companies argued that they were making safe and responsible choices for users. In some cases, that may be true. Grok hallucinated about the outcome of the 2024 election before the votes were counted.
However, the Gemini incident turned into a tailspin. This led Google to disable Gemini's ability to generate images of people, which is not yet possible in the free version of Gemini. Google called the incident a mistake and apologized for being “out of line.”
Andreessen and Sachs don't think so. Both venture capitalists say Google didn't miss the mark at all, but in fact missed the mark a little too clearly. They considered this a pivotal mask-off moment for Google.
In a February 2024 episode of All In, Sachs said in response to the Gemini incident, “The people running Google AI are secretly bringing in their own preferences and biases, and those biases are extremely liberal.'' ” he said. “Do you think they're going to remove the stigma? No, they're going to make it more subtle. That's what I'm concerned about about it.”
“That's 100% intentional. That's how you can Google a black George Washington,” Andreessen said in a recent interview with Weiss, revisiting the Gemini incident. “This relates directly to Elon's argument, which is that the crux of the issue is that we need to train AI to lie.” [i.e., to produce answers like Gemini’s]”
As Andreessen mentioned, Elon Musk is an outspoken opponent of “woke AI chatbots.” Musk originally founded the well-funded AI startup xAI in 2023 to take on OpenAI's ChatGPT, but Musk said at the time that he was infected with the “wake-mind virus.” . He ultimately created Grok, an AI chatbot with significantly fewer safety measures than other leading chatbots.
“I'm going to start something called TruthGPT, or maximal truth-seeking AI, in 2023, trying to understand the nature of the universe,” Musk said in an interview with FOX.
When Musk launched Grok, Sachs praised his efforts. “By having something like Grok around, you can at least keep OpenAI honest, keep ChatGPT honest,” Trump's AI czar said on a November 2023 episode of All In. Ta.
Now, Musk is doing more than just keeping ChatGPT honest. He funded xAI and raised over $12 billion to compete with OpenAI. He is also suing Sam Altman's startup and Microsoft, which could prevent OpenAI from going commercial.
Mr. Musk's influence with conservative government officials has already shown he can be influential in other areas as well. Texas Attorney General Ken Paxton is investigating a group of advertisers that allegedly boycotted Elon Musk's “X.” Musk previously sued the same ad group, and some companies have since resumed advertising on his platform.
It's unclear what Trump and other Republicans can do if they actually want to investigate OpenAI and Google's AI censorship. It could be a professional investigation, a legal challenge, or just a cultural issue that President Trump could pursue over the next four years. Regardless of the path forward, President Trump's Silicon Valley advisers are not mincing words on the issue today.
“Elon used Twitter files to do a privatized version of what should be widely done today,” Andreessen told Weiss, referring to Musk's Twitter censorship allegations. “We, the American people, need to know what's going on during this time, especially the intertwining of government pressure and censorship… There need to be consequences.”