Google’s fiasco with its artificial intelligence program is emblematic of an issue simmering at the tech giant. It’s not a new story. It dates to—you guessed it—the 2016 election. The rollout of Gemini has been something of a trainwreck. Its image creator refused to generate any images of white people, whereas the AI prompt couldn’t say with certainty that Adolf Hitler was a worse individual than Elon Musk. What’s worse, AI—Hitler killing millions in the Holocaust or Musk posting memes? It couldn’t answer, which led to data crunchers, like Nate Silver at FiveThirtyEight, calling for Gemini's shutdown. CEO Sundar Pichai apologized for the controversy, saying, “I know that some of its responses have offended our users and shown bias – to be clear, that’s completely unacceptable, and we got it wrong.”
Says the man who wants to use AI to combat Trumpism.
Michael Shellenberger has a new series out on his Substack that delves into Google, AI, and the tech giant’s campaign to twist public opinion to hurt Republicans in elections (via Public):
…after a Google employee suggested that Trump won due to “misinformation” and “fake news coming from fake news websites being shared by millions of low-information voters on social media,” Pichai specifically pointed to the use of artificial intelligence to achieve the aim of countering “misinformation.”
“I think our investments in machine learning and AI is a big opportunity here,” he said. Machine learning is a form of AI.
Pichai then suggested that Google was already manipulating search results.
“There are many, many places where we are ranking,” he said, “we are algorithmically doing stuff…understanding some of the things that are happening, and course correcting.”
[…]
Pichai emphasized that there would be a technological solution to the information that Google executives believed had resulted in Trump’s victory. “This is a problem of scale and not being able to keep up. Human systems fail in many of these things. So I think investing more in machine learning and AI could be one way we actually make progress.”
Much of the demand for anti-racism comes from within Google. “Speaking to white men,” said a white male Google employee in November 2016, “there's an opportunity for you right now. to understand your privilege in the society. Take the opportunity to go through the Bias-Busting training. Read about privilege. Read about the real history of oppression in our country.”
The Google executive, who nearly started crying, said, “I think the main thing I just wanted to say is give yourself time and space to deal with whatever you're going through. Healing is a process, and it takes time. But one thing that makes Alphabet at Google so special is this term I heard, I'd never heard it before I got here, which is, this is a place where you can bring your whole self to work.”
And like Twitter, Google has worked extensively with the government, earning numerous contracts with the intelligence community. Twitter and the FBI created a robust operation of public opinion influence and censorship, which was exposed in the Twitter Files that Shellenberger, Matt Taibbi, and others wrote about at length. It was no different at Google, where most of its government work was kept secret until a whistleblower came forward. As you can guess, a lot of their government contracting work dealt with peddling propaganda:
Google receives billions of dollars from American citizens. Google split a $9 billion Defense Department contract with other Big Tech firms in 2023 and a multi-billion contract with the CIA. Google got $7 million in 2010. The National Science Foundation and the Department of Defense (DOD) are funding various censorship tools for Google.
Between 2016 and 2020, Google won more than 250 contracts with the military and federal law enforcement agencies. “We remain committed to partnering with the government on projects that are consistent with our terms of service, acceptable use policies, and AI Principles,” said a Google spokesman.
[…]
Much of Google’s work for the government is secret. Google provided the technology for Project Maven, a DOD initiative to build artificial intelligence to track moving targets for drones. But the funding was hidden. None of Project Maven's contracts mentioned Google at all. It was only through employee whistleblowing and investigative journalism that Google’s involvement became known.
Google’s work with Maven was orchestrated through a subcontract with ECS Federal. This U.S. defense contractor provides technology services to arms of the DOD and various federal agencies.
Google misrepresented its work on Maven, reporter Lee Fang discovered. A Google executive claimed the Maven contract was small, at just $9 million, but internal Google emails revealed that Google anticipated the military drone AI revenue would be $250 million per year. “Indeed,” noted Fang, “one month after news of the contract broke, the Pentagon allocated an additional $100 million to Project Maven.”
Former Google CEO and board member Eric Schmidt is a key link between Silicon Valley and the US intelligence community. Schmidt chaired the National Security Commission on AI, the Defense Innovation Board, and the National Security Commission on Emerging Biotechnology. Schmidt’s new investment fund, America’s Frontier Fund (AFF) said in a leaked announcement that it would receive government funding to invest in tech companies.
It’s hard to find the line between Google and the US government. “From inception,” notes [Nafeez] Ahmed, “Google was incubated, nurtured, and financed by interests that were directly affiliated or closely aligned with the US military intelligence community: many of whom were embedded in the Pentagon Highlands Forum.”
[…]
Government censorship and disinformation leaders like the people who started Google Jigsaw call their work “prebunking,” but in many ways it’s just old-fashioned government propaganda. The word “prebunking” hijacks people’s thinking in the same way that “misinformation” does. And, most importantly, “prebunking” is a demand that both social media platforms and news media organizations censor information deemed to be “misinformation.”
The word “prebunking” is itself a kind of misinformation. Another term for “prebunking” is “government propaganda.” But when you call it “prebunking” it sounds like it comes from people who care about the truth, as opposed to propaganda, which rightly has connotations of totalitarianism. But they are one and the same thing.
This is not the first time that the US intelligence and security agencies have been caught engaging in disinformation campaigns aimed at delegitimizing people.
What blocks proper dissemination of these revelations and ensuing accountability is the fact that a) the media is biased, b) there are still significant numbers of Americans—liberals—who still believe in the Russian collusion hoax, and c) these people control the flow of information. Since the Department of Justice decided to toss their highly trusted source, Alexander Smirnov, under the bus for lying to his handlers about Burisma executives bribing the Bidens, the red scare antics have been amplified on the airwaves. The timing of this indictment also isn’t lost on anyone. Smirnov being charged has already re-ignited the Kremlin scares, with additional layers of insanity being added to this fake news cake, like how Russia is using the GOP House majority to interfere in the 2024 elections. It’s all a convenient diversion for the actual election meddlers—the FBI, CIA, and the ‘Big Tech’ overlords who help them control the narrative.
The calls of election interference have been coming from and remain inside the house.