創(chuàng)新工場創(chuàng)始人兼CEO、人工智能工程院院長李開復博士日前在紐約時報（The New York Times）的“觀點”（Opinion）欄目發(fā)表專欄文章《人工智能對人類社會的真正威脅》（The Real Threats of Artificial Intelligence），解讀了上述問題。

在這篇文章里，李開復博士討論了人工智能技術(shù)未來發(fā)展所帶來的幾個更真切和亟待解決的問題：全球性的失業(yè)問題及可能產(chǎn)生的全球性經(jīng)濟失衡和貧富差距。

在2011年紐約時報的一篇專欄文章中，李開復博士曾預測中國手機支付直接跳過信用卡取代現(xiàn)金而成為主流支付方式、線上教育會進入二三線城市、本地餐飲使用發(fā)生顛覆。當時，微信還沒有出現(xiàn)，而如今我們回顧這三個預測，基本都已經(jīng)發(fā)生。

因此我們有理由相信，這次的預測同樣并非科幻小說式的想象，文中的觀點值得深思。

以下為專欄文章的完整中譯版及英文原文：

面對呼之欲出的人工智能時代，您最擔心什么？ 

通常，人們對于這個問題的回答很像各類科幻片中的驚悚情節(jié)。他們擔心人工智能的發(fā)展會帶來所謂的“奇點”，即在人類發(fā)展的某一特定歷史時刻，人工智能會完全超越人類智慧，繼而將人類社會帶入一場無法想象的變革當中。人們甚至開始懷疑，人工智能是否最終會控制人類，使人類淪為所謂的“機械人”。

這些問題值得探討，但并非亟待解決。先不論這些問題是否會發(fā)生，即使哪天真的出現(xiàn)，也是數(shù)百年以后。而目前，人類還沒有任何已知的途徑和方法能夠?qū)斍白钭吭降娜斯ぶ悄芟到y(tǒng)——比如剛剛戰(zhàn)勝了最出色的人類棋手柯潔的圍棋計算機程序AlphaGo，轉(zhuǎn)化為通用的人工智能，即具有自我意識、可進行常識性推理、能夠自覺地從多領域獲取知識、并具有感知、表達和理解等能力的電腦程序。

但這并不意味著我們就可以高枕無憂。恰恰相反，現(xiàn)有人工智能技術(shù)和產(chǎn)品的發(fā)展速度之快大大超出我們的認識和預期，人工智能技術(shù)注定會改變我們的世界，并不完全以我們的意愿為轉(zhuǎn)移。人工智能是工具，不是一種智慧形式。但它注定會重新定義工作的意義以及財富的創(chuàng)造方式；值得注意的是，它將帶來前所未有的經(jīng)濟失衡現(xiàn)象，甚至改變?nèi)虻臋?quán)力格局。

因此，當務之急，讓我們先對這些迫在眉睫的現(xiàn)實挑戰(zhàn)予以關注。 

人工智能到底是什么？粗略來講，人工智能技術(shù)指的是獲取某一領域（比如貸款償還記錄）的海量信息，并利用這些信息對具體案例（是否應給某人貸款）做出判斷，以達成某一特定目標（貸方利益最大化）的技術(shù)。這些技術(shù)在給定任務中所展現(xiàn)出的工作能力已經(jīng)被證明可以完全超越人類的表現(xiàn)。

今天，這樣的人工智能技術(shù)正在被廣泛應用于各個領域。隨著它的進一步發(fā)展，會不可避免地對就業(yè)造成沖擊。很多崗位和職業(yè)會逐步消失，例如銀行出納員、客戶服務代表、電話銷售員、股票和債券交易員等；甚至律師助理和放射科醫(yī)生這樣的工作也會被這類軟件所取代。假以時日，人工智能技術(shù)還會學會控制如無人駕駛汽車和機器人這類半自主或全自主硬件設施，逐步取代工廠工人、建筑工人、司機、快遞及許多其他職業(yè)。

與工業(yè)革命及信息革命不同，人工智能技術(shù)所帶來的沖擊并非單純指向某些特定崗位和職業(yè)，如傳統(tǒng)制造業(yè)中的手工藝者被流水線工人所取代；或只會使用紙張和打字機的秘書被精通電腦的個人助理所替代等；人工智能所帶來的是對現(xiàn)有職業(yè)和工作版圖大規(guī)模地顛覆。毋庸諱言，其中大部分為低薪工作，但某些高薪崗位也將面臨挑戰(zhàn)。

值得注意的是，這場變革將會為開發(fā)人工智能技術(shù)及采用人工智能技術(shù)的公司和企業(yè)帶來巨額利潤。試想，如果優(yōu)步能全面利用無人駕駛車進行運營；蘋果公司能夠省卻大量人力生產(chǎn)其產(chǎn)品；全年滿足超過三千萬筆貸款請求卻不需要任何人工干預的借貸公司；可以想見，這些企業(yè)將利用人工智能技術(shù)創(chuàng)造何等驚人的利潤和收益！而這一切已經(jīng)是現(xiàn)在進行時。創(chuàng)新工場最近就在國內(nèi)投資支持了一家利用人工智能技術(shù)進行借貸的的初創(chuàng)企業(yè)。

誠如你所看到的，人類正面臨著很難妥善共存的兩個發(fā)展前景：一方面我們迎來了僅用少量人力就能創(chuàng)造巨大財富的發(fā)展時代，而另一方面，大量人員也將因此而下崗和失業(yè)。各種權(quán)衡，何去何從？

答案之一當然是教育，即要對人工智能所不擅長的領域進行有針對性的人員教育和再培訓。具體來說，人工智能并不擅長需要創(chuàng)造力、規(guī)劃能力以及“跨領域”思考能力等類型的工作——比如辯護律師。這些能力也是目前很多高端職位所要求的，問題是通過短期培訓來傳授和獲取這些能力和技能的可能行較低。另一個方向則是彌補人工智能系統(tǒng)所欠缺的“人際交往能力”，發(fā)展出更多類似社會工作者、酒保、按摩技師等需要人際間微妙互動的崗位。即便如此，另一個問題隨之出現(xiàn)：我們的社會對酒?；蝾愃茘徫挥钟卸啻笮枨竽?？

按照我的個人推測，要解決人工智能變革所帶來的大規(guī)模失業(yè)問題，需要的是更多我所說的所謂“關愛服務”。 這是人工智能無法完成，而社會又大量需要的服務；更不用講你我生而為人所賴以的使命感和榮譽感。此類服務崗位不勝枚舉，例如：陪伴老人就醫(yī)的志工、孤兒院的教導員、戒酒互助社的志愿者，甚或未來可能出現(xiàn)的——幫助那些沉迷于電腦虛擬現(xiàn)實刺激中的“平行人”重返人生現(xiàn)實的熱心人。換言之，當下的很多所謂志愿服務工作未來都可能成為真正的職業(yè)。 

其中一些服務甚至會轉(zhuǎn)變?yōu)楦咝铰殬I(yè)并趨于專業(yè)化，例如可協(xié)助和配合“人工智能癌癥診斷程序”工作的、具有專業(yè)醫(yī)療知識、同時又富有同情心和極強溝通技巧的醫(yī)療服務提供者?？傮w而言，人們可以選擇比現(xiàn)在更短的工作時間。 

那么，誰會為這些工作買單呢？文章開始時我提到的那些集中于相對少數(shù)企業(yè)手中的巨額財富現(xiàn)在可以派上用場了。在我看來，人工智能所創(chuàng)造財富中的相當一部分會不可避免的轉(zhuǎn)移到那些工作被取代了的人們那里去。而這一過程似乎只能是通過凱恩斯主義的財政政策——即提高政府相關領域的開銷，及增加高利潤公司的稅收來加以實現(xiàn)。 

至于那樣狀況下的社會福利是何種形式，我認為可能是一種有條件的全民基本收入方案，即社會福利將面向有經(jīng)濟需求并符合條件的人群。所謂“條件”，是指福利申請者必須努力參與就業(yè)或再就業(yè)培訓，或保證參與一定工時的“關愛服務”。 

當然，為了給這類社會福利提供資金，提高稅率可能在所難免。政府不僅要補貼大部分人的生活和工作，還要設法對此前大量下崗員工無法繳納的個人所得稅進行彌補。 

這就帶來了關于人工智能最終、也是最重要的挑戰(zhàn)。我所描繪的凱恩斯主義的財政政策或許在美國和中國是可行的，因為這兩個國家可以通過其規(guī)模巨大且成功的人工智能企業(yè)來獲取稅收，并以此支撐其高昂的社會福利方案。但是其它國家又當如何呢？ 

相較而言，其他國家會面臨兩個難以克服的問題。首先，大部分人工智能所創(chuàng)造的財富會流入美國和中國。人工智能是一個“強者更強”的產(chǎn)業(yè)：數(shù)據(jù)越多，產(chǎn)品越好；產(chǎn)品越好，所能獲得的數(shù)據(jù)就更多；數(shù)據(jù)更多，就更吸引人才；人才越多，產(chǎn)品就會更好。在這個良性循環(huán)里，中美兩國目前已經(jīng)匯聚了大量人才、市場份額以及能夠調(diào)動的數(shù)據(jù)。

舉例來說，中國的語音識別企業(yè)科大訊飛以及人臉識別公司如曠視科技、商湯科技等就市值來講，都已經(jīng)成為行業(yè)翹楚。在谷歌、特斯拉及優(yōu)步等企業(yè)的引領下，美國的無人駕駛技術(shù)也是首屈一指。而在消費互聯(lián)網(wǎng)領域，中美七家企業(yè)——谷歌、臉書、微軟、亞馬遜、百度、阿里巴巴、騰訊——都已在其現(xiàn)有產(chǎn)品和服務中大量使用人工智能技術(shù)，并正快速將其運營版圖擴展到全球范圍內(nèi)，盡可能占據(jù)更大份額的人工智能市場。從目前的情勢看，美國似乎占據(jù)發(fā)達國家市場及部分發(fā)展中國家市場，而中國公司無疑贏得了多數(shù)發(fā)展中國家市場。

對于中國和美國以外的其他國家來講，另外一項挑戰(zhàn)則在于許多國家還在日益增長的人口，尤其是一些發(fā)展中國家。龐大的人口可以成為一種經(jīng)濟資本，就如同其近幾十年來在中國和印度的經(jīng)濟發(fā)展中所產(chǎn)生的積極作用。但是在人工智能時代，這一資本卻可能成為經(jīng)濟負擔，因為其中大部分人口將面臨下崗失業(yè)。

所以，如果很多國家不能通過向高額盈利的人工智能企業(yè)征稅來補貼工人，他們還能有什么其他選擇？依我個人推論，為避免本國人民陷入貧困，這些國家會與提供最多人工智能軟件的國家——中國或者美國——進行磋商和談判，最后以特定人工智能企業(yè)在本地用戶中的盈利來換取國家所需的社會福利補貼。從而最終成為中美兩國的經(jīng)濟依附體，這樣的經(jīng)濟發(fā)展態(tài)勢也將重塑當今的地緣政治版圖。

一言以蔽之，最大程度地縮小人工智能可能造成的經(jīng)濟失衡和貧富差距，已是當下必須要考慮的問題，此差距不僅體現(xiàn)在國家內(nèi)部，也體現(xiàn)在國與國之間。從樂觀的角度看：人工智能為我們展現(xiàn)了一個打破全球經(jīng)濟失衡狀態(tài)的機會，而挑戰(zhàn)所帶來的巨大影響，將使任何國家都無法置身事外。

紐約時報上刊登的《The Real Threats of Artificial Intelligence》一文，紅框標示。

英文版：

What worries you about the coming world of artificial intelligence?

Too often the answer to this question resembles the plot of a sci-fi thriller. People worry that developments in A.I. will bring about the “singularity” — that point in history when A.I. surpasses human intelligence, leading to an unimaginable revolution in human affairs. Or they wonder whether instead of our controlling artificial intelligence, it will control us, turning us, in effect, into cyborgs.

These are interesting issues to contemplate, but they are not pressing. They concern situations that may not arise for hundreds of years, if ever. At the moment, there is no known path from our best A.I. tools (like the Google computer program that recently beat the world’s best player of the game of Go) to “general” A.I. — self-aware computer programs that can engage in common-sense reasoning, attain knowledge in multiple domains, feel, express and understand emotions and so on.

This doesn’t mean we have nothing to worry about. On the contrary, the A.I. products that now exist are improving faster than most people realize and promise to radically transform our world, not always for the better. They are only tools, not a competing form of intelligence. But they will reshape what work means and how wealth is created, leading to unprecedented economic inequalities and even altering the global balance of power.

It is imperative that we turn our attention to these imminent challenges.

What is artificial intelligence today? Roughly speaking, it’s technology that takes in huge amounts of information from a specific domain (say, loan repayment histories) and uses it to make a decision in a specific case (whether to give an individual a loan) in the service of a specified goal (maximizing profits for the lender). Think of a spreadsheet on steroids, trained on big data. These tools can outperform human beings at a given task.

This kind of A.I. is spreading to thousands of domains (not just loans), and as it does, it will eliminate many jobs. Bank tellers, customer service representatives, telemarketers, stock and bond traders, even paralegals and radiologists will gradually be replaced by such software. Over time this technology will come to control semiautonomous and autonomous hardware like self-driving cars and robots, displacing factory workers, construction workers, drivers, delivery workers and many others.

Unlike the Industrial Revolution and the computer revolution, the A.I. revolution is not taking certain jobs (artisans, personal assistants who use paper and typewriters) and replacing them with other jobs (assembly-line workers, personal assistants conversant with computers). Instead, it is poised to bring about a wide-scale decimation of jobs — mostly lower-paying jobs, but some higher-paying ones, too.

This transformation will result in enormous profits for the companies that develop A.I., as well as for the companies that adopt it. Imagine how much money a company like Uber would make if it used only robot drivers. Imagine the profits if Apple could manufacture its products without human labor. Imagine the gains to a loan company that could issue 30 million loans a year with virtually no human involvement. (As it happens, my venture capital firm has invested in just such a loan company.)

We are thus facing two developments that do not sit easily together: enormous wealth concentrated in relatively few hands and enormous numbers of people out of work. What is to be done?

Part of the answer will involve educating or retraining people in tasks A.I. tools aren’t good at. Artificial intelligence is poorly suited for jobs involving creativity, planning and “cross-domain” thinking — for example, the work of a trial lawyer. But these skills are typically required by high-paying jobs that may be hard to retrain displaced workers to do. More promising are lower-paying jobs involving the “people skills” that A.I. lacks: social workers, bartenders, concierges — professions requiring nuanced human interaction. But here, too, there is a problem: How many bartenders does a society really need?

The solution to the problem of mass unemployment, I suspect, will involve “service jobs of love.” These are jobs that A.I. cannot do, that society needs and that give people a sense of purpose. Examples include accompanying an older person to visit a doctor, mentoring at an orphanage and serving as a sponsor at Alcoholics Anonymous — or, potentially soon, Virtual Reality Anonymous (for those addicted to their parallel lives in computer-generated simulations). The volunteer service jobs of today, in other words, may turn into the real jobs of the future.

Other volunteer jobs may be higher-paying and professional, such as compassionate medical service providers who serve as the “human interface” for A.I. programs that diagnose cancer. In all cases, people will be able to choose to work fewer hours than they do now.

Who will pay for these jobs? Here is where the enormous wealth concentrated in relatively few hands comes in. It strikes me as unavoidable that large chunks of the money created by A.I. will have to be transferred to those whose jobs have been displaced. This seems feasible only through Keynesian policies of increased government spending, presumably raised through taxation on wealthy companies.

As for what form that social welfare would take, I would argue for a conditional universal basic income: welfare offered to those who have a financial need, on the condition they either show an effort to receive training that would make them employable or commit to a certain number of hours of “service of love” voluntarism.

To fund this, tax rates will have to be high. The government will not only have to subsidize most people’s lives and work; it will also have to compensate for the loss of individual tax revenue previously collected from employed individuals.

This leads to the final and perhaps most consequential challenge of A.I. The Keynesian approach I have sketched out may be feasible in the United States and China, which will have enough successful A.I. businesses to fund welfare initiatives via taxes. But what about other countries?

They face two insurmountable problems. First, most of the money being made from artificial intelligence will go to the United States and China. A.I. is an industry in which strength begets strength: The more data you have, the better your product; the better your product, the more data you can collect; the more data you can collect, the more talent you can attract; the more talent you can attract, the better your product. It’s a virtuous circle, and the United States and China have already amassed the talent, market share and data to set it in motion.

For example, the Chinese speech-recognition company iFlytek and several Chinese face-recognition companies such as Megvii and SenseTime have become industry leaders, as measured by market capitalization. The United States is spearheading the development of autonomous vehicles, led by companies like Google, Tesla and Uber. As for the consumer internet market, seven American or Chinese companies — Google, Facebook, Microsoft, Amazon, Baidu, Alibaba and Tencent — are making extensive use of A.I. and expanding operations to other countries, essentially owning those A.I. markets. It seems American businesses will dominate in developed markets and some developing markets, while Chinese companies will win in most developing markets.

The other challenge for many countries that are not China or the United States is that their populations are increasing, especially in the developing world. While a large, growing population can be an economic asset (as in China and India in recent decades), in the age of A.I. it will be an economic liability because it will comprise mostly displaced workers, not productive ones.

So if most countries will not be able to tax ultra-profitable A.I. companies to subsidize their workers, what options will they have? I foresee only one: Unless they wish to plunge their people into poverty, they will be forced to negotiate with whichever country supplies most of their A.I. software — China or the United States — to essentially become that country’s economic dependent, taking in welfare subsidies in exchange for letting the “parent” nation’s A.I. companies continue to profit from the dependent country’s users. Such economic arrangements would reshape today’s geopolitical alliances.

One way or another, we are going to have to start thinking about how to minimize the looming A.I.-fueled gap between the haves and the have-nots, both within and between nations. Or to put the matter more optimistically: A.I. is presenting us with an opportunity to rethink economic inequality on a global scale. These challenges are too far-ranging in their effects for any nation to isolate itself from the rest of the world.