TED技术:Garry Kasparov:不要畏惧智能机器,与它们合作吧!


TED技术:Garry Kasparov: Don't fear intelligent machines. Work with them

This story begins in 1985,when at age 22,I became the World Chess Championafter beating Anatoly Karpov.Earlier that year,I played what is calledsimultaneous exhibitionagainst 32 of the world'sbest chess-playing machinesin Hamburg, Germany.I won all the games,and then it was not consideredmuch of a surprisethat I could beat 32 computersat the same time.To me, that was the golden age.


Machines were weak,and my hair was strong.


Just 12 years later,I was fighting for my lifeagainst just one computerin a matchcalled by the cover of "Newsweek""The Brain's Last Stand."No pressure.


From mythology to science fiction,human versus machinehas been often portrayedas a matter of life and death.John Henry,called the steel-driving manin the 19th centuryAfrican American folk legend,was pitted in a raceagainst a steam-powered hammerbashing a tunnel through mountain rock.John Henry's legendis a part of a long historical narrativepitting humanity versus technology.And this competitive rhetoricis standard now.We are in a race against the machines,in a fight or even in a war.Jobs are being killed off.People are being replacedas if they had vanished from the Earth.It's enough to think that the movieslike "The Terminator" or "The Matrix"are nonfiction.

There are very few instances of an arenawhere the human body and mindcan compete on equal termswith a computer or a robot.Actually, I wish there were a few more.Instead,it was my blessing and my curseto literally become the proverbial manin the man versus machine competitionthat everybody is still talking about.In the most famous human-machinecompetition since John Henry,I played two matchesagainst the IBM supercomputer, Deep Blue.Nobody remembersthat I won the first match —



In Philadelphia, before losing the rematchthe following year in New York.But I guess that's fair.There is no day in history,special calendar entryfor all the peoplewho failed to climb Mt. Everestbefore Sir Edmund Hillaryand Tenzing Norgaymade it to the top.And in 1997, I was stillthe world championwhen chess computers finally came of age.I was Mt. Everest,and Deep Blue reached the summit.I should say of course,not that Deep Blue did it,but its human creators —Anantharaman, Campbell, Hoane, Hsu.Hats off to them.As always, machine's triumphwas a human triumph,something we tend to forget when humansare surpassed by our own creations.

Deep Blue was victorious,but was it intelligent?No, no it wasn't,at least not in the way Alan Turingand other founders of computer sciencehad hoped.It turned out that chesscould be crunched by brute force,once hardware got fast enoughand algorithms got smart enough.Although by the definition of the output,grandmaster-level chess,Deep Blue was intelligent.But even at the incredible speed,200 million positions per second,Deep Blue's methodprovided little of the dreamed-of insightinto the mysteries of human intelligence.

Soon,machines will be taxi driversand doctors and professors,but will they be "intelligent?"I would rather leave these definitionsto the philosophers and to the dictionary.What really matters is how we humansfeel about living and workingwith these machines.

When I first met Deep Bluein 1996 in February,I had been the world championfor more than 10 years,and I had played 182world championship gamesand hundreds of games againstother top players in other competitions.I knew what to expect from my opponentsand what to expect from myself.I was used to measure their movesand to gauge their emotional stateby watching their body languageand looking into their eyes.

And then I sat acrossthe chessboard from Deep Blue.I immediately sensed something new,something unsettling.You might experience a similar feelingthe first time you ridein a driverless caror the first time your new computermanager issues an order at work.But when I sat at that first game,I couldn't be surewhat is this thing capable of.Technology can advance in leaps,and IBM had invested heavily.I lost that game.And I couldn't help wondering,might it be invincible?Was my beloved game of chess over?These were human doubts, human fears,and the only thing I knew for surewas that my opponent Deep Bluehad no such worries at all.


I fought backafter this devastating blowto win the first match,but the writing was on the wall.I eventually lost to the machinebut I didn't suffer the fate of John Henrywho won but diedwith his hammer in his hand.[John Henry Died with a Hammer in His HandPalmer C. Hayden][The Museum of AfricanAmerican Art, Los Angeles]It turned out that the world of chessstill wanted to havea human chess champion.And even today,when a free chess appon the latest mobile phoneis stronger than Deep Blue,people are still playing chess,even more than ever before.Doomsayers predictedthat nobody would touch the gamethat could be conquered by the machine,and they were wrong, proven wrong,but doomsaying has always beena popular pastimewhen it comes to technology.

What I learned from my own experienceis that we must face our fearsif we want to get the mostout of our technology,and we must conquer those fearsif we want to get the bestout of our humanity.While licking my wounds,I got a lot of inspirationfrom my battles against Deep Blue.As the old Russian saying goes,if you can't beat them, join them.Then I thought,what if I could play with a computer —together with a computer at my side,combining our strengths,human intuitionplus machine's calculation,human strategy, machine tactics,human experience, machine's memory.Could it be the perfect game ever played?

My idea came to lifein 1998 under the name of Advanced Chesswhen I played this human-plus-machinecompetition against another elite player.But in this first experiment,we both failed to combinehuman and machine skills effectively.Advanced Chess foundits home on the internet,and in 2005, a so-calledfreestyle chess tournamentproduced a revelation.A team of grandmastersand top machines participated,but the winners were not grandmasters,not a supercomputer.The winners were a pairof amateur American chess playersoperating three ordinary PCsat the same time.Their skill of coaching their machineseffectively counteractedthe superior chess knowledgeof their grandmaster opponentsand much greatercomputational power of others.And I reached this formulation.A weak human player plus a machineplus a better process is superiorto a very powerful machine alone,but more remarkably,is superior to a strong human playerplus machineand an inferior process.This convinced me that we would needbetter interfacesto help us coach our machinestowards more useful intelligence.

Human plus machine isn't the future,it's the present.Everybody that's used online translationto get the gist of a news articlefrom a foreign newspaper,knowing its far from perfect.Then we use our human experienceto make sense out of that,and then the machinelearns from our corrections.This model is spreading and investingin medical diagnosis, security analysis.The machine crunches data,calculates probabilities,gets 80 percent of the way, 90 percent,making it easier for analysisand decision-making of the human party.But you are not going to send your kidsto school in a self-driving carwith 90 percent accuracy,even with 99 percent.So we need a leap forwardto add a few more crucial decimal places.

Twenty years aftermy match with Deep Blue,second match,this sensational"The Brain's Last Stand" headlinehas become commonplaceas intelligent machinesmovein every sector, seemingly every day.But unlike in the past,when machines replacedfarm animals, manual labor,now they are comingafter people with college degreesand political influence.And as someonewho fought machines and lost,I am here to tell youthis is excellent, excellent news.Eventually, every professionwill have to feel these pressuresor else it will mean humanityhas ceased to make progress.We don'tget to choosewhen and wheretechnological progress stops.We cannotslow down.In fact,we have to speed up.Our technology excels at removingdifficulties and uncertaintiesfrom our lives,and so we must seek outever more difficult,ever more uncertain challenges.Machines havecalculations.We have understanding.Machines have instructions.We have purpose.Machines haveobjectivity.We have passion.We should not worryabout what our machines can do today.Instead, we should worryabout what they still cannot do today,because we will need the helpof the new, intelligent machinesto turn our grandest dreams into reality.And if we fail,if we fail, it's not because our machinesare too intelligent,or not intelligent enough.If we fail, it's becausewe grew complacentand limited our ambitions.Our humanity is not defined by any skill,like swinging a hammeror even playing chess.

There's one thing only a human can do.That's dream.So let us dream big.

Thank you.


来自:VOA英语网 文章地址: http://www.tingvoa.com/html/20180903/Garry-Kasparov-Dont-fear-intelligent-machines-Work-with-them.html