{"text":[[{"start":5.2,"text":"The big US technology companies assert that they are the most trustworthy guardians of AI — only they understand every facet of the fast-evolving technology unfathomable to slower-moving politicians, academics and journalists. But to echo the Roman poet Juvenal: who guards the guardians? One of our best hopes lies with the companies’ own employees, who not only understand AI but dare to speak out. "}],[{"start":31.599999999999998,"text":"To the outside world, tech employees may appear overpaid and pampered, fixated on accruing life-changing stock options and “benefitsmaxxing”. But many clearly feel a deep sense of responsibility about the use of powerful technologies and are prepared to flag their concerns. This week, more than 560 Google employees, co-ordinated by researchers at Google DeepMind, signed a letter urging Alphabet’s leadership not to allow the use of the company’s AI tools for classified military operations. “We want to see AI benefit humanity, not being used in inhumane or extremely harmful ways,” they wrote. “We feel that our proximity to this technology creates a responsibility to highlight and prevent its unethical and dangerous uses.”"}],[{"start":77.3,"text":"Employees at Google, Amazon and Microsoft have previously voiced concerns about how their companies’ products might have been used by the Israeli military to target Palestinians during the Gaza war. Whistleblowers, such as Frances Haugen at Facebook, have exposed how some social media companies have engineered addiction. Her testimony to Congress in 2021 helped underpin the recent landmark legal case against Meta and Google that found social media companies liable for providing services that were harmful to children."}],[{"start":108.19999999999999,"text":"The latest letter from Google employees comes amid a bigger dispute between the Pentagon and the AI research lab Anthropic over the use of its technology. In February, Anthropic rejected a revised contract with the Pentagon, fearing that AI might be misused for domestic mass surveillance or the development of lethal autonomous weapons. The Pentagon responded by designating the company as a supply-chain risk, threatening to cut all its contracts with federal agencies. Anthropic is contesting that decision in court. "}],[{"start":137.39999999999998,"text":"The accelerationist wing of Silicon Valley regards attempts to constrain the Pentagon from deploying AI models for “any lawful use” as absurd. In their view, if companies do not want their products used for national security purposes, they should not supply them. Other AI companies, including Google, xAI and OpenAI, are happy to pick up the work. On Monday, Google extended its $200mn contract with the Pentagon to cover classified operations, saying it was proud to support national security."}],[{"start":170.29999999999998,"text":"The Silicon Valley hawks highlight how the US is locked in a tech arms race with China, which has no democratic constraints. “Our adversaries will not pause to indulge in theatrical debates about the merits of developing technologies with critical military and national security applications. They will proceed,” the tech company Palantir argued in a controversial post on X. "}],[{"start":192.2,"text":"These arguments would carry more force if the Pentagon had shown a willingness to learn from mistakes, rather than seeming to gloss over them. Under defence secretary Pete Hegseth, it appears to have slow-rolled an investigation into the bombing of an Iranian school in Minab in February, which killed 168 people, including 110 children. The preliminary evidence suggests the school was struck by a US Tomahawk missile as a result of faulty, possibly AI-assisted, targeting."}],[{"start":222.29999999999998,"text":"Several tech companies, which are currently shedding thousands of workers as they roll out AI agents, have been taking a tougher line towards employee complaints. In 2024 Google fired 50 employees who had protested against the company selling cloud computing services to Israel. Employees at many tech firms are required to sign non-disclosure agreements as well as non-disparagement contracts on leaving."}],[{"start":245.64999999999998,"text":"That makes it all the more commendable that some employees are prepared to risk their careers by raising their concerns. Employees can never expect to dictate their employers’ policies, nor should they. It is up to a company’s board to decide what principles are inviolable and where trade-offs are appropriate. But if tech executives are sincere in their determination to defend the democratic west, they should look to win the argument rather than close down debate. Tech employees intimately understand the capabilities, limitations and flaws of the AI they are developing. If tech bosses cannot persuade them that the technology is being deployed appropriately, they will struggle to convince the rest of us."}],[{"start":291.04999999999995,"text":""}]],"url":"https://audio.ftcn.net.cn/album/a_1777688206_5602.mp3"}