>> NOW, AS EXPERTS RAISE CONCERNS ABOUT THE AI BUBBLE, THEY'RE ALSO SOUNDING THE ALARM ABOUT HOW FAST THE TECHNOLOGY IS ADVANCING.

THE U. S. AND CHINA ARE RACING TO LEAD THE WORLD IN AI AND THERE'S LITTLE SIGN THAT THAT'S SLOWING DOWN.

OUR NEXT GUEST ARGUES THAT INSTEAD OF COMPETING THE TWO SUPERPOWERS SHOULD BE WORKING TOGETHER TO PUT GUARDRAILS ON THE TECHNOLOGY.

AI SCIENTIST AND FORMER PRESIDENT OF GOOGLE CHINA KAI-FU LEE JOINS WALTER ISAACSON TO DISCUSS HOW THE TECHNOLOGY COULD TRANSFORM BUSINESSES AND RESHAPE FUTURE EMPLOYMENT FOR BETTER OR WORSE.

>> THANK YOU, BIANNA.

AND DR.

ARE KAI-FU LEE, WELCOME TO THE SHOW.

>> THANK YOU, WALTER.

>> IN YOUR VERY TIMELY BOOK "AI NATIVE" YOU HAVE A SENTENCE I'M GOING TO READ TO YOU WHICH IS "THIS IS A TRANSFORMATION FAR MORE PROFOUND THAN THE INDUSTRIAL REVOLUTION AND ITS IMPACT IS BEING FELT NOT IN DECADES BUT IN QUARTERS. "

WHAT HAS SURPRISED YOU THE MOST IN THE PAST SIX MONTHS?

>> IN THE LAST SIX MONTHS THE IMPROVEMENT OF AI ACTUALLY IMPROVED FASTER THAN BEFORE, NOT SLOWER.

USUALLY AS TECHNOLOGIES IMPROVE THEY REACH SOME KIND OF AN ASIM TOET.

BUT AI IS IMPROVING FASTER FOR SEVERAL REASONS.

ONE IS AI IS LEARNING TO CODE MUCH BETTER THAN PEOPLE.

AND THAT ACCELERATES THE SPEED OF SOFTWARE DEVELOPMENT.

ANY SOFTWARE CAN BE DEVELOPED VERY CHEAPLY.

NO LONGER DO YOU NEED EXPENSIVE SOFTWARE ENGINEERS.

SECONDLY, AI IS LEARNING TO IMPROVE ITSELF.

THIS IS CALLED RECURSIVE SELF-IMPROVEMENT.

AND THAT REMOVES THE HUMAN IN THE LOOP.

YOU NO LONGER NEED 2,000 SCIENTISTS TRYING DIFFERENT IDEAS AND AI IMPROVES ITSELF.

NOW, THE 2,000 SCIENTISTS CAN STILL, YOU KNOW, PUT ICING ON THE CAKE, BUT AI IS NOW LEARNING TO IMPROVE ITSELF.

>> LET ME DRILL DOWN ON WHAT YOU JUST SAID, WHICH IS ITS ABILITY TO REMOVE HUMANS FROM THE LOOP.

WHY IS THAT A GOOD THING?

>> WELL, FOR THE SPEED OF INCREASING INTELLIGENCE IT'S A GOOD THING.

FOR MANY THINGS IN THE COMPANY, IN MY COMPANY, I REMOVED ME AS THE CEO IN THE LOOP.

THAT'S NOT A GOOD THING.

BECAUSE AI CANNOT BE ACCOUNTABLE.

AND AS A CEO, AS A MANAGER YOU WILL NEED A PERSON TO TAKE ACCOUNTABILITY.

SO IN THAT SENSE AI CANNOT BE ACCOUNTABLE.

>> BUT ISN'T THAT THE PROBLEM WE'RE FACING, WHICH IS THESE AUTONOMOUS AGENTS SUDDENLY ACTING UNACCOUNTABLY?

>> THAT'S WHY THOSE CASES NEED BETTER GUARDRAILS AND ALSO IN MANY CASES THEY SHOULD NOT BE --THEY SHOULD HAVE A HUMAN IN THE LOOP.

WHEN I TALKED EARLIER ABOUT AI SELF- IMPROVEMENT, THAT'S PURELY BASED ON THE METRIC OF ITS INTELLIGENCE IS INCREASING WITHOUT AS MUCH HUMAN INVOLVEMENT AND IT'S MAKING ITSELF SMARTER.

AND THAT I THINK IS NOT A DIFFERENCE FROM MAKING DECISIONS THAT COULD LOSE TONS OF MONEY OR BREAK A LAW.

>> YOUR BOOK IS A REALLY USEFUL GUIDE FOR HOW SHOULD I APPLY AI TO MY BUSINESS, TO MY LIFE, TO MY CORPORATION .

TELL ME, WHAT ARE COMPANIES GETTING WRONG ABOUT HOW THEY APPLY AI TO THEIR BUSINESS?

>> MOST CEOs STILL THINK ABOUT AI AS A TECHNOLOGY OR A SOFTWARE AND THEY DELEGATE TO THE CIO TO IMPLEMENT A MODEL OR A CAPABILITY OR TO IMPLEMENT AGENTS FOR DIFFERENT DEMTS SUCH AS CUSTOMER SERVICE OR LEGAL OR FINANCE.

NOW, THESE AGENTS ARE USEFUL IN AND OF THEMSELVES, BUT THEY ARE NOTHING MORE THAN SOME COST REDUCTION.

BUT I THINK IF WE THINK ABOUT AI GETTING SO MUCH SMARTER IT'S LIKE HAVING AN IQ300 PERSON ENTER YOUR COMPANY AND HAVING THAT PERSON BE A CUSTOMER SERVICE REP IS REALLY UNDERUTILIZING THE PERSON'S CAPABILITY.

SO IN MY BOOK "AI NATIVE" I TALK ABOUT WHEN WE HAVE AN AI THIS SMART IT SHOULD DO THE MOST MISSION-CRITICAL TASKS IN THE COMPANY SUCH AS INCREASING REVENUE, INCREASING PROFITABILITY, INCREASING GROWTH, INNOVATING NEW PRODUCTS OR HELPING THE CEO MANAGE THE COMPANY BETTER.

SO THESE ARE THE KINDS OF PRODUCTS THAT MY COMPANY ZERO ONE DOT AI DEVELOPS AND I THINK PEOPLE WHO REALLY WANT TO DO AI TRANSFORMATION SHOULD BE THINKING ABOUT HOW THEY CAN USE AI FOR THESE HIGH-VALUE TASKS IN ADDITION TO USING IT FOR SMALL CONVENIENCE OR SMALL COST SAVINGS.

>> WELL, YOU'RE A CEO.

YOU'VE GOT YOUR OWN COMPANY.

TELL ME WHAT YOU DO AND GIVE ME AN EXAMPLE EVEN TODAY, WHAT ARE YOU GOING TO DO TO HAVE AI AS A CEO OF YOUR BUSINESS CHANGE THINGS?

>> RIGHT.

I'M STILL THE CEO.

I MAKE ALL THE DECISIONS.

BUT I RELY ON AI TO GATHER ALL THE INFORMATION, PROVIDE ME WITH INPUT, INSIGHT AND RECOMMENDATIONS.

AND THAT BECOMES POSSIBLE BECAUSE AI CAN READ ALL THE DATA IN MY COMPANY, ALL OF THE SALES REPORTS, ALL OF THE CRM, ERP.

IT CAN READ ALL THE DATA BASES.

IT CAN KNOW EVERYBODY'S WORK PROGRESS REPORT, EVERYBODY'S REVIEWS AND OKRs AND KPIs.

AND ALSO IT CAN SIT IN IN ALL THE MEETINGS LIKE VIDEO CONFERENCES ARE BEING RECORDED, SO THAT'S BEING SAVED.

AND USUALLY THAT'S JUST ARCHIVED BUT NOW WE CAN HAVE AI LISTEN TO ALL OF IT.

CUSTOMER CALLS.

IT'S VERY CUSTOMARY TO RECORD THE CALL WITH A CUSTOMER FOR THE PURPOSE OF SERVING CUSTOMERS BETTER.

AND THOSE LOGS CAN BE SAVED.

SO NOW IF YOU THINK ABOUT AN AI ACTING AS MY CHIEF OF STAFF THAT HAS ACCESS TO ALL OF THIS DATA IT'S A GOD'S EYE VIEW TO ALL THE DATA IN THE COMPANY.

I CAN THEN ASK QUESTIONS LIKE OF ALL THE MEETINGS IN MY COMPANY, NOT JUST THE ONES I WAS AT BUT ALL OF THEM, WHAT ARE THE THREE BIGGEST DEBATES THAT ARE HAD THAT I NEED TO KNOW ABOUT OR WHICH ONE OF MY TOP TWO STRATEGIC INITIATIVES ARE POTENTIALLY BEING BLOCKED AND WHAT IS BLOCKING IT AND HOW DO I UNBLOCK IT?

OR WHO ARE ARE THE BEST AI PROGRAMMERS IN MY COMPANY, WHICH ONES ARE THE BEST LEADERS AMONG THEM.

THEN I WANT TO ELEVATE THEM TO MORE IMPORTANT POSITIONS TO MANAGE THESE AI WORKERS.

AND FINALLY I CAN ASK HOW CAN I BE A BETTER CEO, HOW SHOULD I RUN MEETINGS BETTER, AND BASED ON THE COMPANY DISASTER LAST WEEK HOW SHOULD WE REVISE OUR ARE PROCESS TO AVOID A RECURRENCE?

>> AND WHAT CAN IT NOT DO FOR YOU?

>> IT CANNOT MAKE DECISIONS FOR ME.

IT IN FACT CAN MAKE PRETTY GOOD DECISIONS BUT WE CAN'T LET IT MAKE DECISIONS BECAUSE IT CAN'T BE ACCOUNTABLE.

IT WOULD BE --THERE WOULD BE NO ONE --IT'S UNABLE TO TAKE ACCOUNTABILITY.

IT HAS NO FEELINGS.

IT'S NOT FEARFUL OF FAILURE.

IT'S NOT BEING ENTICED BY PASSION AND SUCCESS.

SO THAT DECISION MAKING, THAT BELIEF AND VISION AND FORESIGHT STILL BELONG TO THE HUMAN.

ALSO, AI CAN ONLY MAKE PREDICTIONS ON THINGS THAT HAVE HAPPENED, SIMILAR THINGS HAVE HAPPENED IN THE PAST.

IT CANNOT INVENT A NEW THING.

BASED ON THE HISTORY OF ALL THE MOBILE PHONES 20 YEARS AGO IT COULD NOT COME UP WITH AN iPHONE.

BECAUSE AN iPHONE WAS CREATED BECAUSE OF STEVE JOBS' VISION, FORESIGHT AND JUST BELIEF THAT THIS IS THE RIGHT PRODUCT AND THAT NO ONE WILL TELL HIM I WANT AN iPHONE BUT ONCE HE BUILDS ONE EVERYONE WANTS ONE.

SO THAT'S THE KIND OF CASE WHERE AI CANNOT PREDICT BECAUSE IT DOESN'T HAVE ANY DATA TO SUGGEST THAT AN iPHONE AS IT WAS DESIGNED WOULD BE A GOOD PRODUCT.

>> YOU HAVE A UNIQUE PERSPECTIVE.

YOU WERE BORN IN TAIWAN, RIGHT?

EDUCATED IN THE UNITED STATES.

YOU SPENT YEARS WITH AMERICAN TECHNOLOGY COMPANIES IN CHINA.

WHEN WE'RE TALKING ABOUT AI'S DEVELOPMENT NOW, ALL THE PEOPLE WE TALK ABOUT ARE KIND OF WORRIED.

THEY SAY WE NEED GUARDRAILS.

PEOPLE SAYING MAYBE WE SHOULD HIT THE PAUSE BUTTON.

AND THEN THEY SAY BUT WE CAN'T DO THAT BECAUSE THEN CHINA WILL BEAT US.

GIVE ME YOUR PERSPECTIVE ON THAT.

>> I THINK CHINA IS DOING A GREAT JOB AS THE ANDROID OF THE AI ECOSYSTEM.

U. S. IS THE iPHONE EQUIVALENT.

SO THE SAME WAY THAT iPHONE LED THE WAY, ANDROID FOLLOWED CLOSELY, iPHONE MADE THE MONEY AND ANDROID HAD MORE MARKET SHARE, THAT'S THE PATH WE'RE ON.

SO I THINK -- >> DOES THAT MEAN THE CHINESE ONE WILL BE OPEN, WHICH IS KIND OF WHAT ANDROID IS, A BIT MORE OF AN OPEN SOURCE?

>> YES.

MOST CHINESE MODELS ARE OPEN SOURCE.

AND THAT'S WHY THEY'RE BEING ADOPTED WIDELY.

OPEN SOURCE HAS THE ADVANTAGE OF BEING CHEAPER AND SOMETIMES EVEN FREE IF YOU ALREADY HAVE THE HARDWARE.

AND ALSO IT HAS THE ADVANTAGE THAT YOU DON'T HAVE TO SEND YOUR DATA TO THE CLOUD BECAUSE YOU CAN RUN IT ON YOUR SERVER OR THE PC.

SO THE CHINESE COMPANIES ARE OBVIOUSLY INTERESTED IN ADVANCING THEIR MODELS.

BUT I THINK THE CHINESE COMPANIES ARE VERY WILLING TO SIT DOWN AND DISCUSS WITH THE AMERICAN COMPANIES WHETHER A DUAL SLOWDOWN IS NEEDED AND CAN WE DO IT ON EQUAL TERMS WITH MUTUAL RESPECT.

I'VE SEEN SOME OF THE AMERICAN EXPERTS SUGGEST THAT AMERICA AND ITS ALLIES SHOULD COME UP WITH THE RULES AND TELL CHINA TO FOLLOW.

I THINK FRANKLY THAT IS VERY ARROGANT AND WILL NOT BE ACCEPTED WELL.

I THINK U. S. AND CHINA ARE THE TWO LEADERS, TO RESOLVE THIS ISSUE TO PROVIDE AI SAFETY THE TWO LEADERS MUST SIT DOWN ON EQUAL TERMS.

JUST LIKE IF PHONES HAVE SOME DANGER AND ROIFD AND iPHONE ARE THE TWO COMPANIES THAT HAVE TO --TWO PRODUCTS, APPLE AND GOOGLE HAVE TO SIT DOWN AND RESOLVE IT.

IT CAN'T BE THE CASE THAT GOOGLE FINDS ALL OF ITS PARTNERS, MAKES THE RULES AND TELL APPLE TO FOLLOW SUIT AND VICE VERSA.

>> WELL, THIS SEEMS VERY INTERESTING, ALMOST NEWSY.

YOU'RE SAYING THAT YOU THINK THAT THE CHINESE ARE READY TO SIT DOWN AND FIGURE THIS OUT.

BUT I COULDN'T THEY FROM YOUR ANSWER WHETHER YOU'RE TALKING ABOUT A GOVERNMENT TO GOVERNMENT THING LIKE WHEN PRESIDENT TRUMP MEETS XI AND HAVING A GOVERNMENT RULES OF THE ROAD, ARMS CONTROL NEGOTIATION, OR WHETHER IT SHOULD BE AS YOU SAID AT THE BEGINNING CHINESE COMPANIES SITTING DOWN WITH AMERICAN COMPANIES AND DOING IT WITHOUT GOVERNMENT.

>> I THINK BOTH ARE VIABLE AND I THINK AMERICANS LEADERS TALK ABOUT BOTH.

I THINK SOME ADVOCATES A PRIVATE SECTOR SELF-RESTRAINT, OTHERS BELIEF REGULATIONS ARE NEEDED.

AND I THINK THIS DANGER IS POTENTIALLY SERIOUS AND NOT FAR FROM IMMINENT THAT BOTH PATHS SHOULD BE PURSUED.

I THINK THE CHINESE COMPANIES ARE MORE COMMERCIAL.

THEY ARE FARTHER AWAY FROM REACHING THIS A. G. I. BUT I DON'T SEE ANY REASON WHY THEY WOULDN'T PARTICIPATE.

I THINK ANY ASSUMPTION THAT CHINESE COMPANIES ARE INHERENTLY OUT TO BUILD SOMETHING BAD OR REFUSE TO PARTICIPATE IS A VERY NAIVE DEMONIZATION OF CHINESE COMPANIES.

AND IF YOU THINK ABOUT PRESIDENT TRUMP AND PRESIDENT XI, LOOK AT THEIR SPEECHES, PRESIDENT XI HAS MENTIONED MANY, MANY, MANY MORE TIMES AI SAFETY.

SO I ALSO THINK CHINA WOULD BE A WILLING PARTICIPANT.

BUT AGAIN, IT HAS TO BE AT EQUAL LEVELS.

>> ONE OF THE DIFFERENCES BETWEEN CHINA AND THE UNITED STATES IS THERE'S A SURVEY THAT SAYS 84% OF THE PEOPLE IN CHINA ARE EXCITED ABOUT AI.

AND I THINK ONLY 38% IN AMERICA ARE.

YOU'VE LIVED IN BOTH PLACES.

WHY THAT DIVIDE?

>> I DON'T KNOW AS MUCH -- I THINK IN THE U. S. THERE ARE A LOT OF ISSUES THAT ARE BROUGHT IN THE OPEN.

THERE ARE A NUMBER OF VOCAL PEOPLE WHO TALK ABOUT DANGERS OF AI.

AND THE MEDIA I THINK SOMETIMES CORRECTLY REPORTS ISSUES BUT SOMETIMES SENSATIONALIZES.

SO THAT LEADS TO I THINK THE RELATIVELY NEGATIVE VIEW.

AND I THINK THE VIEW ABOUT AI TAKING JOBS AWAY IS MORE PREVALENT IN THE U. S. THAT'S WHY AT VARIOUS GRADUATION COMMENCEMENTS THE STUDENTS DIDN'T WANT TO HEAR THE SPEAKER TALK ABOUT AI.

SO THAT'S MY GUESS ABOUT THE U. S. NOW, IN CHINA IF YOU THINK ABOUT WHAT HAS BROUGHT THE SUCCESS TO CHINA IT'S ABOUT LEARNING TO DO MANUFACTURING, LEARNING TO ACTUALLY ADOPT MOBILE PHONES AHEAD OF U. S. IN WIDE USAGE, LEARNING TO USE THE INTERNET, DEVELOPING CHINA'S OWN INTERNET WEBSITES AND BEING THE FIRST TO USE MOBILE DEVICES AND HAVING A LOT OF CONVENIENCES INCLUDING VERY FAST FOOD DELIVERY, VERY GOOD CAR RIDE SERVICES, VERY GOOD E-COMMERCE, ET CETERA.

SO I THINK THE PEOPLE VIEWS ARE VERY MUCH INFLUENCED BY THEIR OWN EXPERIENCES.

SO A TYPICAL CHINESE 40 OLDER WOULD LOOK BACK ON THE LAST 20 YEARS AND ALL THE MIRACULOUS CONVENIENCES, IMPROVEMENTS AND FINANCIAL BENEFITS FROM TECHNOLOGY.

SO IT'S NATURALLY FOR THEM TO FEEL MORE OPTIMISTIC.

NOW, AS WE SEE MORE ISSUES EMERGE IF THEY EMERGE, OBVIOUSLY THAT NUMBER WILL COME DOWN.

>> LET ME END WITH A PERSONAL QUESTION.

YOU'VE HAD A BATTLE WITH CANCER.

THAT RAISES REALLY TWO QUESTIONS FOR ME.

TO WHAT EXTENT DO YOU THINK THAT WE'VE BEEN MISALLOCATING ALL OF OUR CPUs AND PROCESSING AND IQ POINTS AND ENERGY AND BUZZ TO THINGS LIKE CHATBOTS AND PERSONAL AGENTS AS OPPOSED TO DOING BIG THINGS LIKE CURING CANCER?

AND SECONDLY, SINCE IT'S GOING TO DISRUPT WHAT THE NATURE OF WORK IS, WHAT DID YOU LEARN FROM YOUR CANCER ABOUT WHAT GIVES MEANING TO LIFE?

>> CERTAINLY.

ON THE FIRST QUESTION I DO THINK AI FOR SCIENCE, IN PARTICULAR AI FOR DRUG DISCOVERY IS AN INCREDIBLY IMPORTANT AREA.

I THINK PEMIS HASABAS HAS WON A NOBEL PRIZE AND I'M VERY EXCITED TO SEE HIM DEDICATING HIS CAREER TO TRACKING THIS.

MY OTHER HAT IS THE CEO OF SANOVATION VENTURES.

.

WE'VE ALSO MADE NUMEROUS INVESTMENTS IN AI FOR DRUG DISCOVERY.

SO I DO THINK THERE SHOULD BE MORE RENEWED FOCUS ON THAT.

I THINK WE'RE THE AWAY VERY CRITICAL JUNCTURE WHERE OUR LONGEVITY CAN INCREASE, NOT JUST LIVING LONGER BUT LONGER AND HEALTHIER, AND AI IS A GREAT WAY TO DISCOVER DRUGS AND ALSO MAKE AI MAKE SCIENTIFIC BREAKTHROUGHS THAT CAN HELP WITH OUR HEALTH AND THE WELL- BEING OF THE WORLD.

ON THE SECOND QUESTION WHILE THERE'S STILL SOME DEBATE BUT I THINK MOST PEOPLE RECOGNIZE AI WILL REMOVE SOME JOBS FROM THE WORLD.

AND WHAT I'VE BEEN WORRIED SINCE 2017 WHEN I STARTED WRITING "AI SUPERPOWERS" IS THAT WHEN AI STARTS DOING MORE AND MORE COGNITIVE TASKS AND NUMBER OF JOBS IN VARIOUS CATEGORIES DECLINE THAT'S NOT ONLY GOING TO BE A ISSUE WITH PEOPLE'S INCOME BUT BECAUSE SO MANY OF US BELIEVE THAT OUR WORK IS THE ESSENCE AND MEANING OF OUR LIFE AND WHY WE EXIST, SO I THINK THE SOCIETY HAS TO MAKE CHANGES AND CONCESSIONS ON A LOT OF OUR VALUES.

FIRST I THINK WE DO NEED TO FIND JOBS THAT ARE VERY NUMEROUS FOR PEOPLE TO TAKE.

BUT MANY OF THEM MAY NOT HAVE HUGE ECONOMIC VALUE.

THEY MIGHT BE SERVICE JOBS.

THEY MIGHT BE VOLUNTEER JOBS.

THEY MIGHT BE HOME SCHOOLING.

THEY MIGHT BE GOING TO FOSTER HOMES AND ELDERLY HOMES.

THESE I THINK JOBS CREATE SOCIETAL VALUE.

THEY BRING WARMTH AND CONNECTIVITY TO PEOPLE.

AND THAT'S CLEARLY IMPORTANT IN AN AGE WHERE AI IS MAKING THE WORLD 2X, 3X, 10X MORE MONEY AND WE ENTER THE AGE OF PLENT TOOUD WHEN THERE'S ENOUGH MONEY FOR EVERYONE ON EARTH IF THEY COULD BE DISTRIBUTED.

BUT IF PEOPLE ARE UNHAPPY BECAUSE THEY'RE NOT FEELING THEY'RE CONTRIBUTING TO THIS WORLD, THAT'S A PROBLEM.

SO I THINK WE NEED TO SHIFT FROM ARE THE ONLY WAY TO CONTRIBUTE IS TO MAKE MONEY AND BECOME RICH.

EQUALLY RESPECTABLE SHOULD BE PEOPLE WHO MAKE THE WORLD A BETTER PLACE.

AND THERE SHOULD BE RESPECT FOR SERVICE JOBS, WHICH MIGHT BE THE MOST NUMEROUS JOBS THAT THE MOST PEOPLE WILL TAKE.

THEY CURRENTLY DON'T HAVE THE STATUS OF THE LAWYER, DOCTOR, SOFTWARE ENGINEERS, BUT IN THE FUTURE THERE WILL BE A LOT MORE OF TEACHERS AND NURSES AND VOLUNTEERS AND HOME SCHOOLING TEACHERS, ET CETERA.

SO THEY NEED TO HAVE RESPECT AND SOCIAL STATUS AND THEY NEED TO BE PAID NOT FOR THEIR ECONOMIC CONTRIBUTION BUT FOR THEIR SOCIETAL CONTRIBUTION.

>> DR.

KAI-FU LEE, THANK YOU SO MUCH FOR JOINING US.

>> THANK YOU, WALTER.