今年馬斯克含金量最高、爭議最大、資訊量最炸裂的一次深度訪談，終於完整流出。當地時間7月23日，他做客《經濟學人》90分鐘獨家專訪
徹底推翻了自己過往的AI論調，直面文明變局、中美競爭、財富重構與社會危機，句句顛覆行業認知。
這場訪談徹底告別空泛暢想，拋出精準且激進的時代時間表：
5年AI智力超越全人類總和
10年後金錢失去價值、人類工作全面被替代。
曾經堅定呼籲暫停AI、警示AI滅世風險的馬斯克，如今心態徹底反轉：
AI浪潮已無法阻擋，即便有關停按鈕，也絕不按下，因為AI終將帶領人類進入全民富足時代。
本次訪談暗藏多個行業獨家猛料：
他實錘Anthropic手握全新超強AI模型、隨時官宣發佈，揭秘美國AI監管下架背後真相；同時公開自己的全球AI共治方案，主張納入中國企業、同業互查制衡，打破現有監管邏輯。
更重磅的是
他直言中國終將領跑全球AI，電力優勢、光刻技術突破進度遠超外界預判，美國技術封鎖根本徒勞無功。
除此之外，他首度真誠反思從政經歷，坦言自己過度涉足政壇、本末倒置；
大膽預判歐洲社會隱患，言論火藥味拉滿。
從AI安全、財富分配、全民發錢新政，到文明宿命、時代變局，這場訪談道盡未來十年顛覆性變革。
埃隆・馬斯克 ×《經濟學人》
受訪人：埃隆・馬斯克（Elon Musk）採訪者：扎妮・明頓・貝多斯（Zanny Minton Beddoes，《經濟學人》總編輯）地點：特斯拉德州超級工廠時間：2026 年 7 月 23 日
【訪談開篇】
Zanny Minton Beddoes：埃隆・馬斯克，非常感謝你接受本次訪談。很難不用溢美之詞來介紹你。當下在世所有人裡，你正在比任何人都更主動地塑造人類的未來。
Elon Musk：我只是一直在持續創造、搭建各類事物。我天生就是建造者，我單純喜歡創造東西。
Zanny Minton Beddoes：你確實在持續搭建一切。你開創電動汽車，顛覆太空旅行；如今你擁有前沿 AI 大模型、人形機器人 Optimus。我們不妨從宏觀圖景切入，你眼中十年後的世界會是什麼樣子？
Elon Musk：
大約五年之內，人工智慧的綜合智力水平，將會超越全體人類智力總和。十年尺度來看，人類大機率不再是主導世界的力量。如果 AI 和人類的智力差距，遠大於人類與黑猩猩之間的差距，你很難想像黑猩猩能夠掌控全域。人類本質上只是進化後的靈長類，就在不算久遠的過去，我們還生活在叢林之中。除了 “身為人類本身” 這件事，AI 幾乎所有事情都能做得比人更好。
Zanny Minton Beddoes：你多次警示 AI 存在滅絕人類的風險，你預估機率大約 10%–20%。明知存在毀滅性風險，為什麼還要繼續推進？好比一枚有兩成爆炸機率的火箭，你依舊選擇登上飛船？
Elon Musk：
沒錯，我依然會登上這艘火箭。不存在真正按下暫停鍵的辦法。即便我們主動停下，其他國家、其他團隊依舊會持續推進 AI 研發。
阻止技術浪潮幾乎不可能。風險客觀存在，但我們也將迎來前所未有的富足時代。只要不爆發全球性核衝突，物資、商品、服務將極大充裕。
未來機器人承擔絕大多數體力勞動，AI 承接腦力工作，工作將變成人類的可選活動，不再是生存必需品。
Zanny Minton Beddoes：如果工作不再必須，普通人如何維持生計？你設想的經濟體系如何運轉？
Elon Musk：可以推行全民普惠高收入方案。財政部直接向民眾發放資金支票。很多人會立刻擔憂通膨問題。但大家忽略一點：AI + 機器人帶來近乎無限的商品供給，巨量產出反而會帶來通縮壓力，通膨壓力會被極大抵消。未來人們可以自由選擇愛好，園藝、藝術、博弈、科研，工作會如同業餘下棋，純粹出於興趣，而非謀生手段。商店裡物資充足，你可以自主選擇要不要動手勞作。
Zanny Minton Beddoes：
這套設想十分宏大，但當下最大難題是 AI 安全。你支援怎樣的監管方案？
Elon Musk：
單純依靠政府監管存在天然短板。絕大多數政策制定者缺乏前沿技術認知，很難判斷一款前沿模型是否具備危險能力。
我的提議：
所有頂尖前沿 AI 實驗室建立同業互查機制。各家巨頭定期溝通，任何新大模型正式發佈前，提前一至兩周交給競爭對手開展安全審查。
同行具備足夠技術能力識別漏洞與風險；如果研發方拒不整改，再同步報送中美監管機構。競爭關係反而會倒逼所有人保持審慎，同行具備動機發現對手模型存在的安全隱患。我願意和山姆・奧特曼、Demis Hassabis 等人坐下來認真討論這套框架。
Zanny Minton Beddoes：矽谷各大 AI 巨頭彼此激烈競爭，很多人相互敵視，這套同業審查機制落地難度極高。你樂觀嗎？
Elon Musk：
事關人類文明存續，優先順序高於商業競爭。
人類擁有共同的底線。同時治理不能只侷限美國企業，必須納入中國 AI 研發團隊。中國在人工智慧、機器人領域實力雄厚，很有可能在未來某個時間段成為全球 AI 領跑者。
Zanny Minton Beddoes：你如何看待中國 AI 發展潛力？
Elon Musk：中國擁有決定性優勢 —— 電力供給。按照當前趨勢，2026 年中國總發電量將達到美國三倍。AI 資料中心極度消耗電力，電力就是算力時代的硬通貨。很多人低估中國攻克先進光刻機的速度，他們距離突破先進製程瓶頸，比外界預想更近。我近期仔細研究過月之暗面 Kimi K3 這類模型，進展令人印象深刻。人為設定壁壘無法阻擋中國 AI 產業前進，限制交流只會割裂全球技術生態。
Zanny Minton Beddoes：我們聊聊你參與川普政府、執掌 DOGE 政府效率部的經歷。你投入數十億美元支援川普競選，隨後主導大規模聯邦預算削減，拆解 USAID 對外援助機構。外界批評，海外援助驟減直接引發人道主義災難，不少弱勢群體因此承受苦難。
Elon Musk：不存在任何人因此死亡。USAID 相關職能被移交至美國國務院，我們清理的是大量欺詐、浪費的項目資金。大量 NGO 依賴持續撥款，資金縮減之後自然會大肆渲染負面敘事，試圖爭取撥款恢復。全球還有蓋茲基金會等海量慈善機構可以承接援助缺口。坦白說，我這次深度捲入政治，有些過火，被局勢帶著走了。回頭看，我本應該把更多精力專注於企業本身。
Zanny Minton Beddoes：你多次在 X 平台發表關於英國的悲觀言論，你提出按照當前趨勢，英國未來 20 年內有可能爆發內戰。你最後一次實地到訪英國是什麼時候？
Elon Musk：幾年之前。
Zanny Minton Beddoes：我長期居住在英國，你的描述和現實情況嚴重脫節。大量網民依據你的帖子，形成英國社會秩序瀕臨崩潰的印象。
Elon Musk：我只是基於長期社會資料做出推演。我歡迎你和我一同走訪英國各地，實地觀察基層現狀。我支援邊境管控、安全城市、理性財政這一系列主張，很難說這些屬於極端立場。媒體習慣性把合理訴求貼上極端標籤，這是扭曲敘事。
Zanny Minton Beddoes：外界長期指責你平台言論帶有排外、種族相關傾向，不少評論認定你持反穆斯林立場。
Elon Musk：我伴侶擁有印度血統，我們共同養育多名子女，其中一名孩子以印度知名物理學家命名。我的多家企業高管覆蓋所有人種，不存在種族偏見。我反對的從來不是族群本身。如果外來群體帶入和西方主流文明完全衝突的價值觀、支援暴力犯罪，我會明確抵制。媒體總是簡化、歪曲我表達的完整邏輯。
Zanny Minton Beddoes：我們轉向太空計畫，SpaceX 目標何時實現載人登陸火星？
Elon Musk：理想目標 2028 年完成載人登陸火星。長期目標建立自給自足的火星文明，讓人類成為多行星物種，規避單一星球災難帶來的文明滅絕風險。我們還計畫搭建太空資料中心，為 xAI 人工智慧提供算力支撐，打通太空算力生態。
Zanny Minton Beddoes：如何看待全球地緣衝突，例如俄烏和平談判前景？
Elon Musk：任何可持續的和平協議，都需要考慮俄羅斯合理安全訴求，不能完全忽視領土層面現實得失。單純依靠制裁無法永久解決衝突，談判必須兼顧各方底線。
Zanny Minton Beddoes：最後一個問題，你經常產生一種感受：當下發生在自己身上一切太過荒誕，甚至懷疑我們生活在外星文明搭建的模擬宇宙之中？
Elon Musk：時常有這種想法。我經歷的諸多事件離奇到不像是真實世界會發生的劇情。從機率角度推演，文明發展到能夠建構高保真模擬世界之後，模擬世界數量會遠遠多於原生現實，我們身處模擬環境的機率並不低。但無論現實本質如何，我們依舊要努力把未來建設得更好。
英文原版轉錄文字
Zanny Minton Beddoes:Elon Musk, thank you so much for joining me on The Insider. It’s hard to introduce you without superlatives. You are doing more to build the future than anyone alive right now.
Elon Musk:I’m just doing a lot of building generally. I’m an instinctive builder. I just like building things.
Zanny Minton Beddoes:You pioneered electric vehicles, revolutionized space travel. You have cutting-edge AI models, humanoid robots. Let’s start with the big picture. What will the world look like a decade from now?
Elon Musk:Within roughly five years, artificial intelligence will exceed the combined intelligence of all humans. Ten years out, humans will probably no longer be the dominant force on the planet.If the intelligence gap between AI and humans becomes far bigger than the gap between humans and chimpanzees, it’s hard to see chimps staying in charge. Humans are just evolved primates. Not long ago we lived in jungles.AI will be able to do nearly everything better than humans, except the experience of being human.
Zanny Minton Beddoes:You have warned AI carries a 10–20% risk of human extinction. Knowing that risk exists, why keep accelerating development? It’s like boarding a rocket with a 20% chance of exploding. Would you still get on?
Elon Musk:Yes, I would get on. There is no practical off switch. Even if we paused, others would keep advancing AI. You cannot halt this technological wave.Catastrophic risk exists, but we are also heading toward an age of incredible abundance, absent global nuclear war. Robots and AI will take over nearly all labor, and work will become optional, not a necessity for survival.
Zanny Minton Beddoes:If work is optional, how will ordinary people sustain themselves? What economic system do you envision?
Elon Musk:Something like universal high income. The Treasury can simply issue checks to people.Many immediately worry about inflation. But people miss the other side: AI and robots create near-unlimited supply of goods. Massive production can create deflationary forces that offset inflation pressures.Humans will pursue activities for passion — gardening, art, research, games. Work will be like playing chess, done for enjoyment, not survival.
Zanny Minton Beddoes:That is a sweeping vision. Let’s turn to AI safety and regulation. What framework do you support?
Elon Musk:Government regulation alone has inherent limits. Most policymakers lack deep technical expertise to judge whether a frontier model poses severe danger.My proposal: frontier AI labs implement a competitor mutual review system.Before launching new large models, share them with rival labs for one or two weeks of safety auditing. Competitors have the technical capability to spot risks. If developers refuse to address hazards, escalate to regulators in China and the United States.Commercial rivalry gives each firm an incentive to keep others honest. I am open to sitting down with Sam Altman, Demis Hassabis and others to formalize this.
Zanny Minton Beddoes:Leading AI firms are fierce rivals, many openly hostile to one another. Is this mechanism realistic?
Elon Musk:Civilizational risk outweighs corporate competition. We share a fundamental baseline interest in human survival.Also, global governance cannot only include American companies. Chinese AI developers must be part of the framework. China is very strong in AI and robotics, and may become the global leader at some point.
Zanny Minton Beddoes:What do you make of China’s AI trajectory?
Elon Musk:China has a decisive advantage: power generation. By 2026, China’s power output could reach three times that of the United States. AI data centers consume enormous power, and power is the core currency of computing.Many underestimate how quickly China will solve advanced lithography challenges. They are closer than most observers think.I have studied models like Kimi K3 from Moonshot; progress is impressive. Artificial barriers cannot slow China’s AI advancement. Fragmenting global technical exchange only hurts everyone.
Zanny Minton Beddoes:Let’s discuss your work in the Trump administration leading the DOGE initiative. You invested heavily in his campaign, then pushed sweeping federal spending cuts and dismantled USAID. Critics argue sudden aid cuts triggered humanitarian suffering overseas.
Elon Musk:Zero people died as a result. USAID functions were transferred to the State Department. We targeted waste and fraudulent funding.NGOs dependent on grants naturally push negative narratives to restore funding. Organizations like the Gates Foundation exist to fill humanitarian gaps. Frankly, I got carried away by politics. In hindsight, I should have focused more on my companies.
Zanny Minton Beddoes:You repeatedly posted warnings that Britain risks civil war within 20 years under current trends. When was the last time you visited Britain in person?
Elon Musk:A few years ago.
Zanny Minton Beddoes:I live in Britain full-time. Your online portrayal diverges sharply from reality. Millions form bleak views of British society based on your posts.
Elon Musk:My assessment comes from long-term social trend data. I invite you to tour Britain with me and see conditions on the ground.I advocate secure borders, safe cities, sensible fiscal policy. These are not extreme positions. Media often labels mainstream pragmatic views as extremist to shape narratives.
Zanny Minton Beddoes:Critics accuse your commentary on X of carrying anti-immigrant, racial undertones, and many view you as anti-Muslim.
Elon Musk:My partner is of Indian descent, and we share children, one named after a famous Indian physicist. My companies employ people of all ethnic backgrounds. I hold no racial prejudice.I oppose value systems incompatible with Western norms, and I oppose violent crime. Media constantly oversimplify and distort my full arguments.
Zanny Minton Beddoes:Turning to space: when does SpaceX aim to land humans on Mars?
Elon Musk:Targeting 2028 for crewed Mars landing. The long-term goal is a self-sustaining Martian civilization, making humanity a multiplanetary species to avoid extinction risks confined to one planet.We also plan space-based data centers to supply computing power for xAI, building an orbital computing ecosystem.
Zanny Minton Beddoes:What is your view on geopolitical conflicts such as negotiations over Ukraine?
Elon Musk:Any lasting peace deal must consider Russia’s legitimate security concerns. You cannot ignore territorial realities. Sanctions alone cannot resolve conflict; negotiations require compromise from all sides.
Zanny Minton Beddoes:One final question. You have said events in your life feel so surreal you sometimes wonder if we live inside a simulation built by an alien civilization.
Elon Musk:I have that thought often. Many things I have experienced feel too bizarre to belong to baseline reality. Statistically, once civilizations can build high-fidelity simulations, simulated worlds vastly outnumber base reality. The odds we are in a simulation are not negligible.Regardless of the nature of reality, we still have a responsibility to build a better future. (財經會議圈)