去年10月底，中國國家主席習近平在一次重要會議中強調要推動新一代人工智能健康發展；今年2月份，美國總統特朗普也簽署了行政令，欲推動美國人工智能發展。隨著這些重量級國家領導人的關注，人工智能成為一個佔據頭條的熱詞。
Chinese President Xi Jinping stressed boosting healthy development in a new-generation of Artificial Intelligence (AI) in late Oct, 2018. U.S. President Donald Trump signed an executive order launching the American AI Initiative in Feb this year. With the attention of these heavyweight leaders firmly on AI, it has become a buzzword that continues to dominate headlines everywhere.
人工智能將如何改變中國的製造工業？人工智能又能應用在別的什麼領域？中國和美國在這個領域會成為競爭對手嗎？隨著人工智能的發展，未來的人類社會將是什麼樣子？
How will AI change China's manufacturing industry? Where else will it be used? Are China and the U.S. rivals in the area of AI? What happens when people use these advancements?
這些問題都能在China Focus對創新工場董事長李開復的採訪中找到答案。
These are all questions China Focus put to Kai-Fu Lee, chairman of Sinovation Ventures.
中國在發展人工智能方面有何優勢？
What Advantages Does China Have in Developing AI?
李開復：我認為中國在發展人工智能領域方面的優勢非常多，中國擁有廣闊的市場、豐富的大數據資源、堅韌執著的企業家以及眾多出色的風險投資人，他們有充足的資金來維持企業的良好運轉。中國幅員遼闊，人口龐大，積累了豐富的數據資源，因此，中國很多企業都可以運用人工智能來開發有價值的應用程序。
Kai-Fu Lee: I think China simultaneously has a large market and extremely rich data resources, tenacious entrepreneurs and lots of good VCs (venture capitalists) with lots of money. China has a lot more data which allows the entrepreneurs to build valuable applications using AI to optimize these applications.
▲ 在北京，一名美團外賣配送員正在進行外賣派送工作。美團網是當前中國最大的電商服務平臺之一。（圖/中國日報）
A deliveryman for Meituan-Dianping, China's largest on-demand service platform, delivers food in Beijing. (Photo/China Daily)
比如，美團網看上去只是在做外賣業務，和人工智能沒什麼關聯，但實際上他們擁有廣泛的用戶群體，通過人工智能他們能更直接地瞭解到用戶想吃些什麼。
For example, a company like Meituan in China, they are a food delivery company, which doesn't sound like AI but once you deploy and have a lot of users, you actually realize that you need AI to understand and figure out what people want to eat.
就這樣，通過人工智能長期的數據採集和持續的計算分析，美團網的每份外賣運單在成本上節約了70美分。這家傳統互聯網公司，通過合理運用人工智能來降低成本，從而實現了利潤最大化並迅速發展成了“超級獨角獸”企業。我認為他們上市後的市值也許會高達500億美元，這將遠超美國的Yelp和Groupon兩個外賣應用軟件。
Through accumulative data collection and computation, over time, they could reduce the cost of delivery to 70 cents per delivery. That allows an internet company, not an AI company, to use AI to maximize profit, minimize delivery costs and all of a sudden, with a large amount of data, become a super unicorn. They will be worth, I think, $50 billion when they go public, much more than the U.S. Yelp or Groupon.
同樣重要的是，企業家們有能力來找到所謂的產品市場定位。歐洲人認為美國人更努力，在美國中部和東海岸生活的人則認為在硅谷人更努力。但與硅谷相比，中國人在工作上比他們都更加努力和堅韌。
Equally important are entrepreneurs who can find the so-called product-market fit. In Europe, people think of the U.S. as hardworking and tenacious, in the east coast or the mid-west of the U.S., people think of Silicon Valley as hardworking and tenacious, but compared to Silicon Valley, China is much more hardworking and tenacious.
▲ 深圳市羅湖區曾經是一個貧窮的村莊，現在到處都是高樓大廈。（圖/中國日報）
Luohu district in Shenzhen, Guangdong Province in southeastern China, which used to be a poor village, is now a flourishing residential area with modern apartment buildings. (Photo/China Daily)
在中國，一個典型的創業公司經常會實行“九九六”或“九九七”的工作制度，這意味著工作時間是早9點到晚9點，一週工作6至7天。我認為中國企業家非常渴望成功的原因有許多，其中很重要的一點是，中國一直是發展中國家，很多企業家的父輩經歷了十代甚至二十代貧窮期，揹負著家族期望的他們，迫切希望改寫自己的命運，因此他們決心放手一搏，在這片領域闖出一方天地。
In China, the typical start-up company is either 9-9-6 or 9-9-7, which means 9 a.m.-9 p.m., six days a week or seven days a week. I think Chinese entrepreneurs are extremely hungry for success for many reasons, one being that because China has been a developing country, there are many entrepreneurs that have been poor for 10 or 20 generations and family expectations and their own desire to prove themselves are tremendous.
另外，馬雲的故事感召了許多人，這不僅是因為馬雲在商業上取得了成功，個性魅力十足，非常擅長演說，更是因為他一開始也只是個平凡的創業者，大學畢業後的他甚至連份在肯德基的工作都得不到，這與他之後所取得了巨大成功構成了戲劇性的反差。因此，他的勵志故事會讓其他創業者也認同創業是個千載難逢的機會，他們能借此幫助家裡擺脫貧困。如果馬雲可以做到這一點，那麼他們也不妨一試。
Alibaba co-founder Jack Ma is a great role model because not only is he successful, charismatic and a great speaker, but also, because he came from humble beginnings. He couldn't even get a job at KFC after college, so other entrepreneurs feel like this is a once in a 20-generation opportunity for them to make it and potentially help their families get out of poverty. If Jack Ma can do it, they feel they should at least have a shot.
人工智能在中國的應用情況如何？
How Is AI Being Used in China?
李開復：醫療保健是人工智能目前最重要的應用領域，也獲得了大家的普遍認可。人工智能常作為輔助工具，運用於放射、核磁共振成像、CT平掃或病理學等醫療領域。因為在這些領域，人工智能可以通過分析患者的病史，來判斷應進行儀器掃描、物理檢查、化學檢測中的哪些醫療項目。
Kai-FuLee：Healthcare is the most important, and I think universally agreed upon, use of AI. It is to be used as a tool to help with radiology, X-rays, MRIs, CTs, or pathology, considering patient history, determining what the scans and exams and blood tests might represent.
未來，人工智能還可以幫助診斷這些病症，預估發病的可能性，判斷過去的治療是否會引發新的症狀，以及估量不同治療方法的成功概率。
In the future, it could also help with diagnoses, the probability of conditions and past treatments and outcomes leading to new symptoms and diagnoses, and various treatments available and the likelihood of success.
▲ 2018年4月3日，廣州某醫院內，患者正使用人工智能機器人進行掛號。（圖/中新社）
People use an AI-powered robot to check in at a hospital in Guangzhou, South China's Guangdong Province on April 3, 2018. (Photo/China News)
比如說，每位醫生在他或她的職業生涯中最多能治療10000名患者，閱讀1000篇論文和500本書。但機器人醫生卻可以採集10億人的醫療記錄，長期積累起龐大的醫學知識庫，而這些能力卻是人類醫生所望塵莫及的。
Each doctor treats about 10,000 patients in his or her lifetime, while also building on the 1,000 papers and 500 books that they read. But an AI device can be trained on a billion people and their medical records, and it can be amassing all of the knowledge that each doctor has into a giant database.
現實中，即使是最好的醫生也無法閱讀完所有的論文或進行過全部臨床試驗，而患者可能會因此錯過使用新藥或通過臨床試驗獲得更新更好的治療方法的機會。然而，人工智能卻可以做到這一點。
AI can learn all the latest treatments. Even the best doctors cannot read all the papers written or be aware of all clinical trials done, thus patients may miss opportunities to use a new drug or a clinical trial.
▲ 2018年，在中國上海舉行的世界人工智能大會上, 參觀者觀看機器人手術系統演示。（圖/新華社）
Visitors watch a demonstration of robot surgery system during the World Artificial Intelligence Conference (WAIC) 2018 in Shanghai, east China. (Photo/Xinhua)
在未來，醫學將變得個性化，每一位患者都會得到個性化的診療。這就好比人們平時瀏覽微博或今日頭條時，可以定製不同的領域，這樣每個人都能看到各自感興趣的內容。但是，目前大多數癌症患者所接受的診療手段還很單一。
In the future, medicine will become personalized, so that each person may get a different treatment. When you open Facebook, everybody looks at a different news feed, we don't all see the same news feed. But when you treat cancer, most people get a standard treatment, overlooking the possible alternate options.
試想一下，就像我們平時進行內容個性化定製一樣，患者在接受治療時，也能根據自己的病情而得到更有針對性的治療方式，這樣該有多好。
But imagine, just like each of us may want to read a different news feed, each of us may have a different type of treatment that is maximally effective.
在人工智能方面，中國可以和美國比肩嗎？
Is China Catching up with the U.S. in AI?
李開復：如果以中國在人工智能領域所創造的市值、稅收和用戶群等標準進行衡量，中國的確在語音識別、機器翻譯、計算機可視化、人工智能銀行和無人機等方面處於領先地位。
Kai-FuLee：In terms of market value, tax volume and user base, China leads in voice recognition, machine translation and computer visualization, and probably in AI banking and drones.
▲ 2018年4月22日，在福建福州舉辦的首屆數字中國建設成果展覽會上，孩子們被展出的機器人吸引。（圖/新華社）
The photo taken on April 22 shows the a group of children fixing theireyes on a robot exhibited at the First Digital China Summit& Exhibition opening in Fuzhou, capital of Fujianprovince on April 22, 2018. (Photo/Xinhua)
中國企業的優勢往往在於編寫代碼，並確保其實現精準快捷，及時解決客戶的問題，在以上這些方面，中國做得比其他任何國家都出色。
Chinese companies are very strong at writing code, making sure it's lean and fast and solves customers' problems, probably stronger than any other country.
美國做得最好的是人工智能在B2B領域的應用，即企業使用人工智能來提高銷售額，提高員工參與度和客戶滿意度，協助管理人力資源等等。在自動駕駛汽車、機器人等方面也領先一籌。
The U.S. certainly leads in autonomous vehicles and robotics, and it is the strongest in B2B AI, that is, using AI in enterprise applications to help companies improve sales, manage customer and employee satisfaction and human resources.
▲ 自動駕駛概念車型NIOEVE 2020年將在美國量產。
Autonomous vehicles concept model NiO EVE 2020 to be batch production in the United States.
簡單來說，在全球人工智能領域，中美雙方勢均力敵，並遙遙領先於其他國家。若全球範圍內任何一個國家或企業在技術上有所突破，當前人工智能格局或將重新洗牌。我們無法預測哪個國家或哪家企業何時能實現突破，但可以確定的是，在過去的五到七年裡，人工智能領域確實沒有取得新突破。
So roughly speaking, it is about 50-50 right now. But if there is a breakthrough, I think everything gets reshuffled. So whichever country or company makes that breakthrough, it will change the game. But we can't predict when breakthroughs will come. I can say for the last five to seven years, there have been no other breakthroughs.
有些人可能會對此感到困惑，也許他們認為：“嘿，人工智能一直在更新換代，並正快速地趕超人類。我經常能看到 ‘人工智能可以在辯論中擊敗人類’ 或 ‘人工智能在醫學圖像分析中比醫生更厲害’ 等新聞報道。你怎麼說沒有突破呢？”
Some people may be confused because they think, hey, I see a breakthrough every week, there is AI beating humans at Go, at debating, in medical image analysis. What do you mean there are no breakthroughs?
▲ 在湖北省襄陽市平安銀行襄陽支行內, 一個人工智能機器人正在接待顧客。（圖/中國日報）
An artificial intelligence-powered robot welcomes a customer in the Xiangyang branch of Ping An Bank in Xiangyang, Hubei province in central China. (Photo/China Daily)
但實際上，這些有突破性的應用只是根據一篇具有突破性意義的研究論文開發出來的，這篇論文的作者是三位研究人員Jeff Henson，YoungLakon和YasuoBengio，他們撰寫了一系列關於深度學習的論文，目前很多應用都是基於此開發出來。
These are merely applications built on the same research paper by Jeff Henson, Young Lakon and Yasuo Bengio. The three researchers wrote a set of papers on deep learning that was the one single breakthrough on which all the applications are built.
因此，如果中國繼續保持人工智能的良好發展態勢，也將有很大機會在該領域躍居第一，或者保持第二名的位置，總之實力非常強大。
So, if this continues, I think China will have a very good chance to become number one, or at least a very strong number two in the field of AI.
中國人工智能發展前景如何？
What Does the Future Hold for AI and China?
李開復：在接下來的10到15年裡，我可以非常自信地說，中美兩國將成為世界上最偉大的人工智能超級大國。據預估，人工智能將會創造17萬億美元的超級市場價值，也將在減輕貧困和飢餓方面發揮重要作用 。
Kai-FuLee：I think in the next 10 or 15 years, I can say with great confidence that China and the U.S. will be the two greatest AI superpowers in the world. AI will deliver tremendous value, it will create about $17 trillion in the next 10 years. So, the impact of AI, in terms of reducing poverty and hunger, will be tremendous.
未來30到40年，人類所有常規性的工作都將由人工智能取代，人們將從這些單調重複性的工作中解放出來。當你下班回家感到渾身疲憊時，你可能不再需要親自下廚準備今天的晚飯；當你去商店購物時，你所需要的一切都將由人工智能機器人自動傳送給你；如果你趕時間想吃快餐解決這一頓飯，那你可以在路上就點完餐，到店裡就能吃上準備好的食物，而這一切都可以由機器人完成。
Eventually, in 30 or 40 years, all routine jobs will be done by AI. People will be incredibly liberated because they no longer have to cook dishes. In stores,everything you need will automatically be sent to you and if you want fast food, it will be instantly prepared and available when you arrive.
▲ 肯德基上海一家分店的百度智能機器人可以與客戶交談, 接受食品訂單。（圖/中國日報）
A Baidu Deur assistant at a KFC outlet in Shanghai can talk with customers and take food orders. (Photo/China Daily)
試想一下你有一半的時間得到解放，不需要受無意義的重複性工作牽絆，而可以專心做你喜歡的事情，並花更多時間陪伴你愛的人們，這樣人類會有更高的幸福感。如果接下來的40年裡，我們能努力規避人工智能帶來的風險與挑戰，這些場景真的會實現。
Imagine half of your time being freed up, not having to do routine work and being able to do things that you love and spend more time with the people you love. I think that is the likely scenario in 40 years.
話雖如此，接下來的20到25年裡我們的確會面臨很多挑戰。比如，現在每個國家都有不同類型的法律，這使獲取廣泛的數據變得十分困難。
But having said that, it will be a rather challenging in the next 20-25 years because currently, every country has different kinds of laws that makes access to large amounts of data more difficult.
例如，全球最大的高質量癌症數據庫只有兩萬個樣本。但對於計算機的可視化技術，一個人臉識別器要接受50億個樣本的訓練。看看這兩者的差距吧，我們還有很長的路要走。我是癌症倖存者，因此，我知道90％的癌症患者都很樂意提供他們的治療相關數據，為人工智能研究出一分力。
For example, the largest high-quality cancer database only has 20,000 samples, but for computer visualization, a face recognizer is trained on 5 billion samples, so look at that gap. Thus, I think we still have a long way to go. As a cancer survivor myself, I personally think 90 percent of cancer patients would happily donate their data for the purpose of AI research.
▲ 浙江省瑞安市，在溫州瑞明工業股份有限公司的廠房內，工業機器人正在搬運生產材料。（圖/新華社）
Industrial robots move production material in a factory of Wenzhou Ruiming Industrial Co.Ltd in Ruian. (Photo/Xinhua)
另外，就目前來看，很多人的工作都是常規性的。如果人工智能取代了他們賴以生存的工作，這些人是否能接受這一情況，仍是個問題。我們有責任為這個行業發展做出更穩妥的規劃，為全人類建設一個更美好的未來。
But governments don't seem to allow this to happen. There are many people whose jobs are largely routine and those will be over time taken by AI. There are people whose job is the meaning of their lives and I think they can't accept such a utopian answer, but I also think we need to be responsible and help plan a good future for all of us.
在未來，一切皆有可能。雖然人工智能的應用確實存在挑戰，需要我們花一定時間努力去克服。但在未來15到20年裡，我確信通過使用人工智能，人類的整體健康水平將得到顯著提高，平均壽命也會得以延長，人們也許會因而對人工智能有所改觀。
So, the possibilities are endless, but I think there are challenges in the application of AI that will take time. Still, if we look at a 15-20-year horizon, I do think our overall health and longevity will be significantly enhanced through the use of AI.
撰文 / China Focus團隊
圖片 / 網絡
閱讀更多 ChinaFocus聚焦中國 的文章
