Protesters March in San Francisco Urging Halt to AI Development
A significant assembly of demonstrators convened in San Francisco, the nucleus of artificial intelligence (AI) innovation, on Saturday. Their resounding call: to compel companies driving AI advancements to “cease the AI race.”
Approximately 200 individuals brandished an array of placards adorned with slogans such as “stop slop,” “it’s not too late to regulate,” and “in a race off a cliff, no one wins.”
They traversed the streets from OpenAI to Anthropic and Google DeepMind, imploring the chief executives to collectively suspend all new AI model training.
Protesters vociferously criticized AI’s purported contributions to escalating housing costs, workforce displacement, environmental degradation, and existential threats confronting future generations.
The rally was orchestrated by Stop the AI Race, under the aegis of activist and ex-AI researcher Michaël Trazzi.
Last year, Trazzi initiated a hunger strike outside Google DeepMind in London to spotlight the imperative for a cessation of AI development activities. A concurrent hunger protest unfolded outside Anthropic’s locale on Howard Street.
Participants, encompassing students, AI professionals, and long-time San Francisco residents, expressed aspirations that their consolidated efforts would heighten public consciousness and consequently amplify pressure on AI corporations.
As they marched through the downtown area, observers from restaurants and balconies paused to capture the unfolding scene.
“Protests can only do so much,” remarked Aleesa Carbo, a Johns Hopkins University student currently involved in AI research. “But if we can raise public awareness, that can mobilize them to communicate their concerns to their senators and the government regarding AI companies.”
Carbo, engaged in the Machine Learning Alignment & Theory Scholars (MATS) fellowship—an esteemed program leading graduates to establish AI safety entities or secure positions at firms like Anthropic—has shifted her focus towards AI safety.
“I am not fundamentally opposed to AI, yet I believe the companies’ current trajectory towards its development is not conducted with due responsibility,” she articulated.
“Ultimately, these models are black boxes. Even we, the practitioners involved in their training and experimentation, possess only a limited understanding of their inner workings.”
Michaël Trazzi, the figurehead of Stop the AI Race, addresses an audience outside OpenAI’s headquarters, invoking the gravity of the situation. “We are in an emergency,” he urged. “They cannot halt the race unless others do.”
Trazzi and his organization are championing a worldwide consensus to pause AI advancements, extending to nations such as China.
Practically, this would entail that existing models remain operational while prohibiting “new training runs of larger or more generalized frontier models.”
As per the Stop the AI Race website, “teams currently enhancing these models’ capabilities would pivot towards narrow AI applications or alignment research.”
As of yet, the targeted companies have not issued a direct response to Stop the AI Race’s appeals. However, the group frequently references a January interview wherein DeepMind CEO Demis Hassabis indicated a potential willingness to advocate for a collective developmental pause.
He envisions an analogous structure to CERN, where leading intellects collaborate to methodically conclude pivotal developments.
“Regrettably,” he added, “international cooperation is required. Even if a single entity or the Western hemisphere undertakes such measures, it lacks efficacy unless a global agreement is achieved.”
Duncan Haldane, the CEO of a San Francisco startup deploying AI in circuit board design, joined the demonstration with his young children in tow.
He aims to sway prominent AI lab leaders towards acknowledging inherent risks and committing to a developmental pause.
Despite acknowledging the advantages AI has conferred upon his enterprise, he characterizes the technology as an “existential threat to humanity.”
“The current capabilities are profoundly perilous and will significantly influence societal structures,” he asserted.
Dean Preston, a former San Francisco supervisor, criticized AI enterprises for their “devastating” impact on the city, notably concerning soaring rents, housing costs, job losses, and environmental chaos. “These tech magnates perceive San Francisco as a mere trophy,” he remarked, “an object to be exploited.”
He cited the public’s resistance against AI developments across California, referencing protests in Pittsburg against a data center by AI developer Avaio Digital and Monterey Park’s decisive ban on data centers, setting a precedent as the first U.S. city to implement such restrictions.
While some protesters yearned for a universal development halt, others called for intensified local regulation of the AI sector.
“I would like our mayor and Board of Supervisors to begin regulating AI within the city,” stated Kathe Burick, a long-term San Francisco resident.
“They should either impose a shutdown or demand a halt, failing which they should restrict operations within city limits.”
Burick equated the reckless advance of AI to the iconic scene from “2001: A Space Odyssey,” wherein astronaut Dave Bowman implores the HAL 9000 computer to open the pod bay doors, receiving the chilling reply, “I’m afraid I can’t do that.”
Reflecting on a prior Stop the AI Race demonstration at OpenAI that garnered a meager turnout of about 30 individuals, Burick expressed optimism regarding the substantial crowd on Saturday, particularly noting the presence of engaged youth.
Source link: Sfchronicle.com.
