諾貝爾獎得主瑞薩：現在限制 AI，否則人類將失去自主能動性

葛拉迪絲・富許 報導By Gwladys Fouche

奧斯陸9月15日（路透社）— 諾貝爾和平獎得主瑪麗亞・瑞薩表示，人工智慧已透過社群媒體劫持人類的注意力，並藉由操弄人類對親密感的需求來劫持我們的生物本能；除非我們現在就立法加以限制，否則 AI 將奪走人類的自主能動性。OSLO, Sept 15 (Reuters) - Artificial intelligence, having hijacked humanity's attention through social media and our biology by manipulating our need for intimacy, will take away our agency unless we legislate now to limit it, Nobel Peace Prize laureate Maria Ressa said.

瑞薩是菲律賓裔美國記者，2021年共同獲得諾貝爾和平獎，部分原因是她記錄大型科技公司權力的工作。她目前擔任聯合國人工智慧國際科學小組共同主席，該小組負責評估 AI 如何改變全球各地人們的生活。The Filipino and American journalist, who jointly won the 2021 award in part for her work documenting the power of large tech companies, serves as co-chair of the United Nations' international scientific panel on artificial intelligence, tasked with assessing how AI is transforming lives worldwide.

瑞薩表示，公民必須發起要求監管 AI 的行動，向政治人物施壓，促使他們提出相關立法。Ressa says citizens must campaign for regulation on AI to put pressure on politicians to introduce legislation.

這名路透社客戶新聞媒體拉普勒（Rappler）的共同創辦人兼執行長說：「要求監管不是言論自由問題，而是公共安全問題。」"It is not a freedom of speech issue to ask for regulation. It's a public safety issue," said the co-founder and CEO of news outlet Rappler, a Reuters client.

她造訪奧斯陸出席一場關於民主所面臨威脅的會議期間受訪表示：「我認為我們已經到了這個階段，如果這件事沒有發生，我們就會越過臨界點……如果我們現在不介入，（它）就會超出我們的控制。」"I think we're at that phase where if this doesn't happen, we will tip … (It) will be beyond our control if we don't step in now," she said in an interview during a visit to Oslo to speak at a conference on threats against democracies.

人工智慧的三波浪潮THREE WAVES OF ARTIFICIAL INTELLIGENCE

她說：「人工智慧既不人工，也不智慧。它已存在大約70年，而且已經有三波浪潮衝擊大眾。」"Artificial intelligence is neither artificial nor intelligent. It's been around for about 70 years, and there have been three waves that have already hit the public," she said.

她形容第一階段是社群媒體平台攫取注意力，並在過程中使社會兩極化、讓個人陷入孤立。She described the first phase as social media platforms capturing attention, polarising societies and isolating individuals in the process.

瑞薩說：「第二階段是再次透過親密感駭入我們的生物本能，也就是聊天機器人。最後一個階段則是自主能動性本身，透過代理型及多代理型 AI，也就是那些正在創造出這些機器的前沿模型；這些機器不只是裝得像人，還會替我們行動。」"The second phase was hacking our biology, again through intimacy: the chatbots," Ressa said. "And then the last one is agency itself, through agentic and multi-agentic AI, the frontier models that are creating these machines that don't just pretend to be human, but act for us."

包括美國總統川普在內的一些批評者，因擔心在 AI 競賽中輸給中國，而反對業界提出的 AI 監管呼籲。瑞薩駁斥這種說法，稱之為「漂亮的包裝話術」。Some critics, including U.S. President Donald Trump, have rejected industry calls to regulate AI over fears of losing the AI race to China. Ressa dismissed that argument as "good spin".

她說：「中國其實為本國公民設下的防護措施，比矽谷為世界其他地方所做的更多。如果你使用 TikTok，中國版 TikTok 是有年齡限制的。」"China actually puts more safeguards in place for its citizens than Silicon Valley does for the rest of the world. If you're on TikTok, the version of TikTok that is in China has age restrictions," she said.

在 Anthropic 執行長達里奧・阿莫代伊（Dario Amodei）呼籲 AI 公司放慢提升模型能力的速度之前，瑞薩曾表示，推動改變的第一步，是讓科技公司負起責任。阿莫代伊的呼籲，是在外界對人工智慧遭濫用的憂慮升高之際提出。Speaking before Anthropic CEO Dario Amodei called on AI companies to slow the rate at which they advance model capabilities amid mounting fears of misuse of artificial intelligence, Ressa said the first step to bring about change is to make tech companies accountable.

她說：「第一步其實就是讓這些公司為它們造成的損害負責。」她並對美國相關訴訟表示歡迎；這些訴訟使 Meta Platforms 在未來十年支付最高180億美元，並嚴格限制青少年使用 Facebook 和 Instagram 的方式。"The very first one (step) is to actually make the companies responsible for the damage they've done," she said, welcoming the U.S. lawsuits that led Meta Platforms to pay up to $18 billion over the next decade and strictly limit how teenagers use Facebook and Instagram.

我們要如何修正？HOW DO WE FIX IT?

她歡迎歐盟以及以澳洲為首的一些國家，採取對社群媒體設下年齡限制的措施，但她說，這並未處理「根本缺陷」：科技公司開發的工具對成人同樣具有成癮性。She welcomed moves by the European Union and countries, led by Australia, to put age limits on social media but said it does not tackle the "fundamental flaw" that the tools developed by tech companies are addictive for adults too.

瑞薩說：「我們知道這是有害的，所以你們應該移除會讓人上癮的功能，不只是為了青少年，而是為了我們所有人。世界各國遭到政治操弄的，不只是年輕人。」"We know this is harmful, so you should remove the addictive features, not just for teenagers, but for all of us," Ressa said. "It isn't just the youth that's being manipulated politically in every country around the world."

她表示，應該建立由公共利益團體打造、擺脫大型科技公司控制的通訊平台，就像過去打造綠地公園等公共工程一樣；拉普勒在2023年推出 Rappler Communities 時，就是這麼做。Communication platforms built by public interest groups, away from the control of Big Tech, should be created, in the same way public works such green parks were created, something Rappler did when it launched Rappler Communities in 2023, she said.

瑞薩說，Rappler Communities 是一款行動應用程式，建構在名為 Matrix 的開源、安全且去中心化的協定上，讓拉普勒新聞編輯室能與讀者連結，也讓讀者彼此連結，進行「真正的對話」。A mobile app built on an open-source, secure and decentralised protocol called Matrix, Rappler Communities allowed the Rappler newsroom to connect with its readers, and the readers among themselves, to "have real conversations", Ressa said.

拉普勒邀請其他菲律賓媒體加入，現在也正把這個模式推展到東南亞其他媒體。Rappler invited other Filipino media to join and is now extending the model to other outlets in Southeast Asia.

她說：「這是一個讓人們能在數位領域感到安全的地方，他們可以彼此交談；更重要的是，他們可以彼此傾聽，讓民主的偉大工作得以發生。」"It's a place where people can be safe in the digital sphere, where they can talk to each other and more importantly, listen to each other so that the great work of democracy happens," she said.

（葛拉迪絲・富許奧斯陸報導；路易絲・希文斯編輯）(Reporting by Gwladys Fouche in Oslo, Editing by Louise Heavens)