自動化是一個“溫柔”的終結者

Barack Obama had a parting shot for his successor this week. A day before Donald Trump predicted he would be “the greatest jobs producer that God ever created”, the outgoing US president appeared like the ghost at the feast, warning of “the relentless pace of automation that will make many jobs obsolete”.

巴拉克?奧巴馬(Barack Obama)給他的繼任者留下了一句臨別贈語。在唐納德?特朗普(Donald Trump)預言自己將成爲“上帝創造的最偉大的就業製造者”一天前，離任的美國總統似乎像是宴會上的幽靈，警告稱“自動化的無情步伐將淘汰很多就業崗位”。

It is at least possible that both will be proved right. Automation has been a constant for decades, and the latest advances in robotics and artificial intelligence all but guarantee that the pace will accelerate. But timing is all. For companies and their investors — no less than for politicians — the key question is not “whether”, but “when”.

這兩種說法至少都可能被證明爲正確。幾十年來，自動化一直是個常數，而機器人和人工智能領域的最新進展基本上確保了其應用速度將加快。但關鍵在於時機。對於公司及其投資者（對於政治人士也是如此）而言，關鍵問題不是“是否”，而是“何時”。

For society at large, the pace of automation will determine how easily the displacement of workers can be handled — and whether the political backlash grows worse. The pace is equally important for the companies trying to push the latest robots and smart machines into the real world, and their investors. Few are in the position of Google parent Alphabet, which has taken the long view on bets like driverless cars — and even Alphabet these days has a new sense of impatience about when it will see returns from “moonshots” like this.

對於整個社會而言，自動化的步伐將決定應對工人失業的難度有多大，以及政治反彈會否加劇。對於那些試圖將最新機器人和智能設備推向真實世界的公司及其投資者而言，自動化的步伐同樣重要。幾乎無人處在谷歌(Google)母公司Alphabet那樣的位置上，該公司對押注於自動駕駛汽車等領域持長遠眼光，但如今就連Alphabet也開始對這樣的“登月式項目”何時帶來回報感到不耐煩。

The variables that will affect the rate of adoption are huge. In a new report on automation this week, McKinsey estimates that half of all the tasks people perform at work could be automated using technologies that have already been proven. But this estimate gives no clue about how long it will take.

影響普及速度的變量很多。在有關自動化的一份最新報告中，麥肯錫(McKinsey)估計，在人們從事的所有工作中，有一半可以通過已得到驗證的技術被自動化。但這個估計沒有說明整個過程需要多長時間。

Given the uncertainties about everything from regulation to the ability of companies to change their processes, the consultants estimate it could take anything from 20 to 60 years. Try building an investment model with that level of variability. Take the case for autonomous cars and trucks. Much of the technology has already been demonstrated, and the potential markets — for both vehicle makers and tech suppliers — are vast. But will it take five, 10 or 30 years for this to become a significant market?

考慮到各種不確定性（從監管到企業改變流程的能力），諮詢顧問們估計，這可能需要20年至60年。試着根據那種程度的變異性來構建一個投資模型吧。以自動駕駛汽車和卡車爲例。很多技術已得到驗證，而潛在市場（對於汽車製造商和科技供應商而言）是巨大的。但是，要讓它變成一個可觀的市場，需要5年、10年還是30年？

Car companies are already spending hundreds of millions of dollars on building driverless car platforms. At this month’s Consumer Electronics Show and Detroit Auto Show, it was clear that driverless technology has graduated from the experimental: carmakers are now racing to bring this technology to the roads. The biggest companies are able to amortise this cost over a large vehicle fleet, but the increasing level of technology in vehicles will challenge many of the industry’s smaller players.

汽車製造商已投入數億美元打造無人駕駛汽車平臺。在本月的拉斯維加斯消費電子展(Consumer Electronics Show)和底特律車展(Detroit Auto Show)上，無人駕駛技術顯然已擺脫實驗階段：汽車製造商正競相把這項技術推向道路。最大的公司有能力將研發成本分攤在較大的車輛保有量上，但汽車的技術含量越來越高，將讓該行業很多規模較小的參與者面臨挑戰。

Companies like Audi talk of autonomy as a progression. It says 60 per cent of new car buyers already opt to pay $3,000-$6,000 for features like automated acceleration and braking. Those customers might reasonably be expected to keep paying up for additional levels of safety and convenience. The shift from cars that stay in their lanes automatically to hands-off-the-wheel driving might turn out to be a smooth — and profitable — evolution.

奧迪(Audi)等公司將自動化稱爲一種進展。該公司表示，60%的新車買家已選擇支付3000至6000美元添加自動加速和剎車功能。可以合理地指望這些消費者爲更高水平的安全和便利付出更高價格。從汽車自動保持在自己的車道上，到開車不用手握方向盤，這一轉變可能會是一場平穩（且有利可圖）的演變。

But the strongest business case for driverless cars comes from the more radical, all-or-nothing step of eradicating the need for human drivers. In a report on automation’s impact on the economy late last month, the White House said that most of today’s 1.7m drivers of heavy trucks in the US are likely to be replaced — though it added that “it may take years or decades” for this to happen.

但無人駕駛汽車的最強大商業理由來自於更爲徹底的、要麼全有要麼全無的一步：消除對人類駕駛員的需要。在上月末一份有關自動化對經濟影響的報告中，白宮表示，目前美國的170萬重型卡車司機中，多數人很可能會被取代，儘管該報告補充稱，這“可能需要幾年甚至幾十年”纔會發生。

There are some very practical considerations. As Michael Chui, a partner at McKinsey, points out it is vastly expensive to replace the estimated 2m heavy trucks on US roads, with an average lifespan of 20 years. Even without new driverless technology, McKinsey estimates it would cost $320bn. But there are likely to be specific investment cases for speedier adoption. Long-haul routes are the low-hanging fruit of trucking. Platooning, in which trucks form a convoy behind a lead truck driven by a human, could bring a form of supervised automation. Although the long tail of automation may take decades, the market for early adopters could still be vast.

有一些非常實際的考量。正如麥肯錫合夥人邁克爾?崔(Michael Chui)指出的那樣，更換目前在美國道路上行駛的大約200萬輛重型卡車成本極其高昂，這些卡車的平均壽命爲20年。麥肯錫估計，即便沒有新的無人駕駛技術，更換成本就將高達3200億美元。但很可能會有加快採用新技術的特定投資理由。長途路線是卡車運輸領域最容易摘取的果實。車輛結隊（Platooning，多輛卡車在由人類駕駛的領頭卡車後面組成一個車隊）可能帶來一種受到監控的自動化形式。儘管自動化的長尾效應可能耗時幾十年，但早期採用者的市場仍可能巨大。

Rather than wiping out jobs immediately, progressive automation might make the lives of today’s truckers more comfortable and then make up for an expected driver shortage in the mid term, before eliminating jobs eventually. This prospect represents the rosy scenario for the companies leading the AI and robotics charge. But as today’s turbulent political climate shows, they would be foolish to count on such a smooth transition.

漸進自動化不會立即消滅就業崗位，而是可能會先讓卡車司機的日子更舒適，然後在中期彌補司機數量的短缺，最終纔會消除這些就業崗位。這種前景爲那些引領人工智能和機器人潮流的公司描繪了一副美好的願景。但就像當今動盪不安的政治氣候所顯示的那樣，指望如此平穩的過渡將是愚蠢的。