America’s artificial intelligence corporate sector is alive and thriving, but many in the tech trenches are expressing mounting concern that President Biden and his team are out to kill the golden goose in 2024.
AI makers are beginning to grumble about Mr. Biden’s sweeping AI executive order and his administration’s efforts to regulate the emerging technology. The executive order issued in late October aims to curtail perceived dangers from the technology by pushing AI developers to share powerful models’ testing results with the U.S. government and comply with an array of rules.
Small AI startups are complaining that Mr. Biden’s heavy regulatory hand will crush their businesses before they even get started, according to In-Q-Tel, the taxpayer-funded investment group financing tech companies on behalf of America’s intelligence agencies.
The concerns come as rivals such as China are racing ahead with their own AI-subsidized sectors, striving to claim market dominance in a technology some say will upend and disrupt companies across the commercial spectrum.
In-Q-Tel’s Esube Bekele, who oversees its AI investments, said at the GovAI conference last month that he is hearing from startups who believe Mr. Biden’s order may create a burden that will disrupt competition and prove so onerous that they cannot survive.
SEE ALSO: Adoption of AI tech a matter of life and death; scenarios show gravity of maintaining human control
“There is a fear from the smaller startups,” Mr. Bekele said onstage at the conference. “For instance, in the [executive order] it says after a certain model there is a reporting requirement. Would that be too much?”
The reporting requirements outlined in the executive order say that the secretaries of Commerce, State, Defense and Energy, along with the Office of the Director of National Intelligence, will create technical conditions for models and computing clusters on an ongoing basis. The order specifies various levels of computing power that require disclosures to the government in the interim until additional technical guidance can be issued.
Mr. Biden said “one thing is clear” as he signed the executive order at the White House: “to realize the promise of AI and avoid the risks, we need to govern this technology.” He called the order “the most significant action any government anywhere in the world has ever taken on AI safety, security, and trust.”
But the R Street Institute’s Adam Thierer said the executive action tees up a turf war in the administration for who is in charge of AI policy. He said he is not yet sure who will prevail, but much AI regulation will now likely happen away from public view.
Mr. Thierer, a senior fellow on R Street’s technology and innovation team, said a tsunami of regulatory activity on AI is incoming.
“So much AI regulation is going to happen off the books,” Mr. Thierer said. “It’s going to be in the so-called ‘soft-law’ arena, soft-power area, through the use of jawboning, regulatory intimidation and sometimes just direct threats.”
SEE ALSO: Hype and hazards: Artificial intelligence is suddenly very real
Concerns across the board
The little guys are not the only ones afraid of Mr. Biden’s regulations and shadow pressure.
Major players in the AI sector such as Nvidia have already gotten an early glimpse of the Biden administration’s plans, and don’t like what they see.
Santa Clara, Calif.-based Nvidia joined the trillion-dollar market capitalization club in May as businesses rushed to acquire its chips for a range of AI technologies involving medical imaging and robotics.
But the Commerce Department’s anticipated restrictions on Nvidia’s sales to China next year, in light of security concerns amid rising tensions between Beijing and Washington, caused the company’s stock to plunge, jeopardizing billions of dollars in expected sales.
Commerce Secretary Gina Raimondo knows her team will cause pain for tech companies and has said she is moving ahead anyway because she believes it is necessary for national security.
Speaking at the Reagan National Defense Forum in California last month, Ms. Raimondo said she told the “cranky” CEOs of chip companies that they may be in for a “tough quarterly shareholder call,” but it will be worth it in the long run.
“Such is life. Protecting our national security matters more than short-term revenue,” Ms. Raimondo said.
Reflecting rising divisions in the exploding marketplace, some tech companies have cheered the push for new regulations, saying they prefer bright legal lines on what they can and cannot do with AI.
Large companies such as Microsoft and Google called for regulations and met with Mr. Biden’s team ahead of the executive order’s release. Microsoft has pushed for a new federal agency to regulate AI.
Several leading AI companies made voluntary commitments to Mr. Biden’s team to develop and deploy the emerging technology responsibly, and some analysts say the established, well-heeled corporate giants may be better positioned to handle the coming regulatory crush than their smaller, start-up rivals.
“The impulse toward early regulation of AI technology may also favor large, well-capitalized companies,” Eric Sheridan, a senior equity research analyst at investment banking giant Goldman Sachs, wrote in a recent analysis.
“Regulation typically comes with higher costs and higher barriers to entry,” he noted, adding that “the larger technology companies can absorb the costs of building these large language models, afford some of these computing costs, as well as comply with regulation.”
But concerns are growing across the AI sector that Mr. Biden’s appointees will look to privately enforce the voluntary agreements.
Mr. Thierer said implicit threats of regulation represent a “sword of Damocles” approach to tech regulation, an approach previously used as the dominant form of indirect regulation in other tech sectors including telecommunications.
“The key thing about a sword of Damocles regulation is that the sword need not fall to do the damage. It need only hang in the room,” Mr. Thierer said. “If a sword is hanging above your neck and you fear negative blowback from no less of an authority than the president of the United States, …you’re probably going to fall in line with whatever they want.”
Mr. Thierer said he expects shakedown tactics and jawboning efforts to be the governing order for AI in the near-term, especially given what he said was dysfunction among policymakers in Washington.
• Guy Taylor contributed to this report.