In a daring gesture last month, Scale AI’s CEO Alexandr Wang took out a full-page advertisement in the Washington Post, urging President Trump that “America must emerge victorious in the AI battle.”
This declaration elicited mixed responses, as demonstrated during Wang’s appearance on the opening night of Web Summit Qatar. When Wang’s interviewer, Felix Salmon from Axios, asked the audience if they supported his view, only two hands were raised. Conversely, Salmon noted a significant majority disagreed.
In response to the dissent, Salmon pressed Wang to justify his stance. “AI is set to fundamentally alter the landscape of national security,” Wang stated. He shared his background growing up in Los Alamos, New Mexico, known as “the birthplace of the atomic bomb,” and mentioned that both his parents were physicists at the National Lab.
Wang views this as a competition between the U.S. and China, expressing his fears that advancements in AI could enable China to “leapfrog” the military capabilities of “Western powers,” a sentiment that inspired his full-page advertisement.
His views resonate with those emerging from defense tech startups and venture capitalists, who advocate for increased autonomy in AI weaponry. They raise concerns that while China might deploy fully autonomous AI weapons, the U.S. could be hindered by the need for human oversight prior to any engagement.
Wang also argued for the necessity of a clear choice between China and the U.S. in terms of foundational large language models (LLMs), suggesting this will be a binary competition, without acknowledging other competitors like France’s Mistral. He posited that U.S. models inherently incorporate principles of free speech, while those from China reflect communist ideological frameworks.
It’s worth noting that researchers have identified many well-known Chinese LLMs exhibit traits of government censorship. Additionally, concerns over potential backdoors for data harvesting by the Chinese government further complicate the trustworthiness of these models.
Wang’s concerns about governmental influence on AI appeared particularly relevant, as his presentation coincided with Scale’s announcement of a collaboration with the Qatari government. On Sunday, Wang revealed that Scale will assist Qatar in developing 50 AI-driven governmental applications spanning sectors such as education and healthcare.
Primarily known for employing numerous contract workers, often located outside the U.S., to aid in model training, Scale collaborates with major players like Microsoft, OpenAI, and Meta, covering most of the leading foundational models in the U.S. It also provides various other products, including an AI data engine and applications tailored for the defense sector.
The distinctly pro-American rhetoric is likely advantageous for Scale AI among its clientele in the Department of Defense. However, the discussions at the Web Summit also highlighted a considerable unease regarding the U.S. attaining AI supremacy.
Compiled by Techarena.au.
Fanpage: TechArena.au
Watch more about AI – Artificial Intelligence
