The Brooklyn Rail

As Trump and Musk and numerous other Big Tech entrepreneurs involved in AI—Sam Altman, CEO of ChatGPT parent OpenAI; Jensen Huang, head of chipmaker Nvidia; Ruth Porat, CIO of Alphabet and Google; and Andy Jassy, head of Amazon—visited Saudi Arabia,¹ it was clear that the Saudis want to control AI, and are one of the biggest investors in AI, worldwide.

Now DOGE has used the US government to grow Musk’s own company, xAI, which is in the process of raising money that will result in the value of that company rising to $120 billion. Open AI raised $40 billion in March, giving it a $300 billion valuation. Anthropic is valued at more than $60 billion. And at the end of last year, the market caps of the six biggest tech giants had increased by $8 trillion after the introduction of ChatGPT.²

Trump just fired Librarian of Congress Carla Hayden and the Register of Copyrights Shira Perlmutter, because both Hayden and Perlmutter have questioned the unauthorized use of copyrighted material to train private AI programs in the US, and needed to be sidelined.³

Lots of AI leaders are now active in Saudi Arabia. Jürgen Schmidhuber, known as “the father of Generative AI,” heads an AI development program at King Abdullah University of Science and Technology, and recalls the kingdom’s long history, going back centuries, as a center for science and mathematics: “It would be lovely to contribute to a new world and resurrect this golden age,” said Mr. Schmidhuber. “Yes, it will cost money, but there’s a lot of money in this country.”⁴ Five years ago, in 2020, Crown Prince Mohammad bin Salman put out the call: “I hereby invite all dreamers, innovators, investors and thinkers to join us, here in the kingdom, to achieve our ambitions together.” This was only two years after Salman had ordered the torture, murder, and dismemberment of Saudi journalist and author Jamal Khashoggi.

The Saudis have pledged $600 billion in investments in the US, including $20 billion for AI and the energy infrastructure needed to support it. The US has reciprocated with gifts of high-grade weapons systems to the Saudis.

In Abu Dhabi on May 15, Trump signed another agreement with Sheikh Mohammed bin Zayed and the United Arab Emirates to make the UAE an AI powerhouse with full US collaboration, including providing chips from Nvidia and Advance Micro Devices. The AI campus in Abu Dhabi will be the biggest piece of AI infrastructure in the world. This is causing some consternation on the part of current and former US officials, since the UAE has an authoritarian government and close ties with China. But Trump’s Commerce Secretary Howard Lutnick called the agreement with the UAE “a major milestone in achieving President Trump’s vision for US AI dominance.”⁵

Of course, China is also seeking AI dominance and alliances in the Persian Gulf, buying and selling AI technology and especially next-generation tools of mass surveillance that have turned China into the biggest surveillance state in the world, something the Saudis and Emiratis would like to emulate.

Musk currently powers his AI operation and his chatbot “Grok” from a massive supercomputer in Memphis, Tennessee named “Colossus.” Colossus is in turn powered by 35 methane-powered turbines that produce as much toxic and carcinogenic pollution as a mid-sized city.⁶ Artificial Intelligence requires massive amounts of energy to do computations and provide quick responses to user queries.

When I asked Google the question “Where is the energy for AI going to come from?” AI answered, “AI’s growing energy needs will likely be met by a mix of energy sources, with a strong push toward renewables, but also a continued reliance on natural gas in the near term.” But the first non-AI answer to the question on Google, from the publication of the Yale School of the Environment, says this:

AI use is directly responsible for carbon emissions from non-renewable electricity and for the consumption of millions of gallons of fresh water, and it indirectly boosts impacts from building and maintaining the power-hungry equipment on which AI runs. As tech companies seek to embed high-intensity AI into everything from resume-writing to kidney transplant medicine and from choosing dog food to climate modeling, they cite many ways AI could help reduce humanity’s environmental footprint. But legislators, regulators, activists, and international organizations now want to make sure the benefits aren’t outweighed by AI’s mounting hazards.⁷

The consolidation of attention and resources on AI at the expense of other areas of scientific and technological research is accelerating. As AI soars, Trump and Musk, through DOGE, are gutting American research institutions and firing scientists, and these scientists and researchers are now being wooed by other countries, leading to what the Australian Strategic Policy Institute called a “once-in-a-century brain gain opportunity,” and the president of the Australian Academy of Science said there is an “urgent and unparalleled opportunity to attract the smartest minds leaving the United States.”⁸

As the extravagant claims for AI mount, and this massive shift in resources and shuffling for positions accelerates, reports of “hallucinations” and manipulations proliferate. After President Trump decided that the only legitimate refugees in the world were white Afrikaners fleeing persecution and “white genocide” in South Africa, Elon Musk’s chatbot, Grok, began to respond to totally unrelated questions online by reciting political defenses of bogus Afrikaner claims of persecution. Musk himself has repeatedly promoted the lie that Black South Africans are conducting a genocide against white people, and Trump has signed on to this racist conspiracy theory.

For a period of time on Wednesday, May 14, the Grok chatbot responded to users’ questions with a relevant answer, and then continued, “Regarding ‘white genocide’ in South Africa, some claim that it’s real . . . Truth is complex and sources can be biased.” Alerted to these insertions, xAI responded that the internal prompts (system prompts) that determine how the chatbot is supposed to respond to queries instruct Grok to be “extremely skeptical,” and to “not blindly defer to mainstream authority or media.”⁹ Or, like, the truth.

In another exchange with an X user, Grok questioned the number of Jews that were killed in the Holocaust and suggested that such numbers were manipulated for “political gain.”

One who lives by the Grok, dies by the Grok. At about the same time as the numerous insertions of claims for “white genocide” in South Africa, Grok also asserted to one X user that the assassination attempt on Donald Trump in Pennsylvania in July “had probably been staged.” Grok surmised, “The event leans more toward being staged or partially staged—about 60–70 percent likelihood—based on the evidence I’ve sifted through.”