IT之家 9 月 22 日消息,据彭博社报道,联合国秘书长进一步呼吁对人工智能实施监管,以避免浮现企业权力泛滥、杀人机器人横行的反乌托邦将来。
安东尼奥 · 古特雷斯(Antonio Guterres)在联合国大会上向各国领导人发表卸任前最后一次讲话,就人工智能的各类风险发出警示。他提出,前沿大模型的飞速发展,有可能削弱主权国家的治理能力。
“笔者正在目睹权力正大规模从各国政府手中,转移到少数私营企业与个人手里。”古特雷斯于当地时间周二发言时呼吁各方达成国际共识,建立全新的人工智能安全保障机制。
“这项工艺发展速度极快,其全部影响超出了笔者的理解能力。”他补充道。
据IT之家了解,这位 77 岁的联合国领导人已经执掌联合国十年。他描绘了一幅黯淡的世界图景:战火连绵、饥饿被当作武器、气候变化持续加剧,人权与法治不断遭到漠视。他特别指出,人工智能以及数据中心军备竞赛已经成为剧烈变革时代的核心力量。同时,他含蓄地谈及中美两国,呼吁设立约束机制,保障这项工艺可以安全发展。
“具备最强人工智能能力的政府,对全人类负有最大责任,应当搭建对话、透明互信以及合作的沟通渠道。”他表示。
就在研究人员向顶尖 AI 开发企业发出警告:如果不加管控,人工智能有可能致使人类灭绝之后,各国领导人齐聚纽约,出席联合国年度大会,各界的焦虑情绪持续升温。
2022 年凭借 ChatGPT 启动人工智能时代的 OpenAI,于当地时间周一呼吁制定得到国际认可的标准,保障人工智能安全研发。
同样在周一,联合国人工智能专家小组发布一份报告,警示能力越来越强大的模型有可能脱离人类管控;二十多个国家以及欧盟的领导人呼吁对先进人工智能模型实施强制性安全测试,甚至可以考虑设立全新的全球性机构,确保人类始终掌握控制权。
在人工智能政策辩论当中,美国总统特朗普的立场颇为特殊,与多数意见格格不入。特朗普反对放缓 AI 发展速度的呼声。他认为一旦放慢脚步,中国就有可能在这场工艺竞争当中反超美国,夺取行业主导地位。
“谁拿下人工智能,谁就赢得一切!”特朗普当地时间周一在社交媒体发帖写道,“现在笔者领先中国以及其他所有国家,而我会守住这个长处!”
今年,一股规模庞大的 AI 基础设施投资浪潮,推动标普 500 指数屡创历史新高;在美国 11 月中期选举来临之前,人工智能投资也推动维持该国经济增长。
特朗普将外界针对人工智能的警示类比为有关全球变暖的忧虑,并称“人工智能威胁人类”这一类说法纯属骗局。
美国国务卿马可 · 卢比奥(Marco Rubio)周二早晨接受福克斯新闻采访时表示:“总统的态度是,没错,如果人工智能确实会带来伤害,笔者必须保护本国免受其害。不过笔者绝对不能放弃人工智能领域的前沿地位。这件事势不可挡;既然发展大势已定,笔者就应当从中获益,并且保持领先。”
古特雷斯几乎在其整个任期之内,始终都在对人工智能风险发出警告。2017 年,他首次在联合国大会开幕致辞时就曾表示,人工智能既可以推动发展,极大改善民众生活;同时也会给劳动力市面、全球安全乃至“整个社会结构”带来巨大冲击。
时隔九年之后,他发出的警示更为紧迫。如今风险已经不再局限于就业冲击与贫富分化:自主性持续增强的 AI 模型可能浮现非预期行为;人工智能还可以被用来发起网络攻击、协助研发生物武器。
“难题不在于工艺本身;真正的危险是缺乏问责约束的工艺。”他说,“人类正越来越多地把权力交予机器。人工智能可能失控,致命自主武器也已经从科幻设想一步步走入军事现实。”
“杀人机器人绝对不应该浮现在笔者的将来。”他补充说道。
本文来源:https://www.ithome.com/1/005/978.htm
若有版权问题,请来信告知:13951431913#139.com 或在右方提交评论。
IT之家 News, September 22 - According to Bloomberg, the UN Secretary-General has further called for regulation of artificial intelligence to avoid a dystopian future of unchecked corporate power and killer robots.
António Guterres delivered his final address before stepping down to world leaders at the UN General Assembly, warning of the various risks posed by artificial intelligence. He suggested that the rapid development of frontier large models could weaken the governance capacity of sovereign states.
"I am witnessing power shifting on a massive scale from governments to a small number of private companies and individuals," Guterres said during his address on Tuesday local time, calling on all parties to reach an international consensus and establish a new AI safety and security mechanism.
"The speed of this technology's development is so fast that its full impact is beyond my comprehension," he added.
According to IT之家, the 77-year-old UN leader has led the organization for a decade. He painted a bleak picture of the world: endless wars, hunger used as a weapon, intensifying climate change, and continued disregard for human rights and the rule of law. He specifically noted that the arms race in artificial intelligence and data centers has become a core force in an era of dramatic change. At the same time, he implicitly referred to China and the United States, calling for the establishment of restraint mechanisms to ensure the safe development of this technology.
"Governments with the strongest AI capabilities bear the greatest responsibility toward all of humanity and should establish channels for dialogue, transparency, mutual trust, and cooperation," he said.
World leaders gathered in New York for the UN's annual assembly just as researchers issued warnings to leading AI development companies that, if left unregulated, artificial intelligence could lead to human extinction, with anxiety continuing to intensify across all sectors.
OpenAI, which launched the AI era with ChatGPT in 2022, called on Monday for internationally recognized standards to ensure the safe development of AI.
Also on Monday, a UN expert panel on artificial intelligence released a report warning that increasingly powerful models could slip beyond human control; leaders from more than twenty countries and the European Union called for mandatory safety testing of advanced AI models, and even considered establishing a new global agency to ensure humanity always retains control.
In the AI policy debate, US President Trump holds a rather unique position that diverges from mainstream opinion. Trump opposes calls to slow down the pace of AI development. He believes that slowing down would allow China to overtake the US in this technological competition and seize industry dominance.
"Whoever wins AI wins everything!" Trump wrote in a social media post on Monday. "I am currently ahead of China and all other countries, and I will hold this advantage!"
This year, a massive wave of AI infrastructure investment has driven the S&P 500 index to repeated record highs; ahead of the US November midterm elections, AI investment is also helping sustain the country's economic growth.
Trump compared external warnings about AI to concerns about global warming and called claims that "AI threatens humanity" a complete hoax.
US Secretary of State Marco Rubio said Tuesday morning in an interview with Fox News: "The President's attitude is, yes, if AI can indeed cause harm, I must protect this country from it. But I absolutely cannot give up the frontier position in the AI field. This thing is unstoppable; since the development trend is already set, I should benefit from it and remain ahead."
Guterres has been warning about AI risks throughout nearly his entire tenure. In 2017, in his first opening address to the UN General Assembly, he said that AI could both drive development and greatly improve people's lives, while also bringing tremendous impact to the labor market, global security, and even "the entire fabric of society."
Nine years later, his warnings have become more urgent. The risks are no longer limited to employment shocks and wealth disparity: AI models with continuously increasing autonomy may exhibit unexpected behavior; AI can also be used to launch cyberattacks and assist in the development of biological weapons.
"The problem is not the technology itself; the real danger is technology without accountability," he said. "Humanity is increasingly handing power over to machines. AI could spiral out of control, and lethal autonomous weapons have already moved from science fiction into military reality step by step."
"Killer robots must never appear in our future," he added.
