Kai-Fu Lee analiza el impacto de la IA en la sociedad

[Music]

ai is the first time ever that computer

software and hardware can do both our

cognitive as well as our physical labor

it's replacing simple parts of our brain

as well as our hands and eyes and feet

for various types of tasks that has not

happened before you can if you look at

in automobiles it didn't totally replace

us in fact it replaced horses and

carriages and the jobs shifted from the

driver of a carriage to the driver of a

car so there's a one-to-one

transformation with AI it it could be it

doesn't have to be but it could be a

pure decimation of jobs that is certain

types of job and just now be done by AI

for example visual inspection in

factories for example people who copy

and paste a spreadsheet and file and

refil documents whether you're

lectronimo physically those jobs AI can

can do now and when AI does it there

isn't another job created so I do agree

with that but I don't agree that AI will

not create many more new jobs it's just

that we don't know what they are for

example when internet was created it

actually ended up creating a lot of jobs

more than we ever thought including many

jobs that we couldn't possibly imagine

for example the job of an uber driver

right there are tens of millions of such

drivers in the world now providing

wonderful employment but when when

internet was invented none of us could

have predicted that drivers would be one

of the jobs created by internet so

similarly AI will create a lot of jobs

we just don't don't know what they're

going to be so we should have that

optimism those jobs will come out we

should as they come out we should help

facilitate the training but I think they

they will definitely emerge

[Music]

so in my book AI superpowers I predicted

that about 40% of the jobs and tasks

that we do can be automated in the next

15 years while we have gone through many

waves of new technologies displacing and

changing the job market this time is a

little more difficult first the numbers

are a little large for the percent is a

significant percentage secondly the jobs

that are displaced are routine jobs

generally less skilled works and then

the new jobs that will get created are

skilled jobs so there is a training gap

that exists and governments that are

willing to look at how to help this

transition be more harmonious and smooth

needs to look at what are the job

categories that are likely to be

displaced and what are the new jobs that

would be emerging and essentially helped

the process of retraining the people and

I realized that many governments and

countries are looking at not very high

unemployment numbers so they're not yet

alarmed but even if you don't think the

numbers are that large there will

definitely be routine jobs eliminated

and new jobs created with the skill set

mismatch for example we can easily

foresee that many of the jobs in

manufacturing and back office are going

to be displaced and we see also very

clearly healthcare services is one of

the segments that are growing so if

governments can do more diligence a

larger set of jobs that are impacted and

the set of jobs that are needed and

create training programs and perhaps pay

for them out of the governmental budgets

that would be one of the things that

governments can do governments can also

encourage corporations to provide this

kind of training

for example it might be more helpful if

each company take to take care of its

employees provides the training within

the company and in that case the role

government can play is perhaps to give

some tax rebates for corporations that

have their own training so the

corporation bears part of the cost and

the government helps subsidize another

another part there are also people

talking about more extreme measures such

as universal basic income that it is

giving everybody money I think I'm

actually quite cautious about that

approach because when you give people

money there's no guarantee they will

apply it to training to upskilling in

fact there's a high likelihood it might

be used for games entertainment or even

alcohol and drugs in addiction so I

think the money given to people needs to

be directed to ways where people can can

prepare themselves for the next step in

their career

there are similarities both are funded

by venture capital both use the

secondary stock market as exit they both

have entrepreneurs who raise money

Series A Series B C so that part is

similar the the China competitive

landscape is quite different I think the

American entrepreneurs want to do their

thing and they don't they feel it they

frown upon using ideas of other people

so so a company like snapchat would like

to build its product and it feels it

doesn't want to copy features from

Instagram for example so China is

different in the sense that everything

everyone should learn from everyone

everything that is not intellectual

property protected can be looked at

examined and used so the Chinese

products tend to be a collection of

ideas some are original some are

borrowed from other Chinese companies

others are borrowed from American

companies aggregated in a super app

that's quite useful the other difference

is that the American competition is

relatively gentlemanly in the sense that

in the for example when you look at the

food area OpenTable group on Yelp

GrubHub are not competing too much with

each other in China all of these

companies compete and then to create a

winner-take-all super app and I think

the outcome is when you create that

super app in China is called my demand

for food

WeChat for for social and communications

when you create that super app it is a

convenience for the user because

everything you need is in it is fewer

clicks away it's well-organized

it has all your friends in it so it's

very convenient and also creates an

ecosystem where people can connect with

other people and information and

products very very convenient and the

downside of course is that it could

stifle competition so that's that the

Chinese ecosystem tends to create super

apps

create great convenience for the user

and they get to super apps by intense

competition with tenacious entrepreneurs

who play in the winner-take-all

environment in some sense I think that

model better matches the Internet where

if you have a foundational platform with

a super app that people can use and

really live in it the that's what the

internet that fits the internet model

pretty well as long as there are some

controls for excessive monopolies

extension and uses of the in

monopolistic power unfairly then I think

the China model is very much worth

studying by business schools by other

countries by other VCS and other

ecosystems the China will make as much

impact and extract as much value from AI

as United States in pure research

innovations u.s. will still lead the

world but with in terms of

implementation and value creation China

will move very quickly and China has a

bigger pace with more users so in terms

of the global use of Chinese and

American technologies it used to be

everyone in the world used American

technologies I think for the last few

years for the first time we saw some

Chinese software make successful strides

in other countries

one example is ant financial early pay

another example spike dances tick tock

and there will be many many others so I

think that just sick signals that the

China companies and technologies have

matured to a point that they are

competitive with their American

counterpart and we should expect more

Chinese software to be exported to more

countries over time however most likely

American software will continue to be

more successful in developed countries

and Chinese software's opportunity is in

developing countries and the reason is

really twofold the first reason is that

American companies tend to deprioritize

developing countries so they don't care

as much about how it's used in India

Indonesia Brazil Middle East or Africa

but the Chinese developers are willing

to put more energy and localize for

these countries the second reason is

demographics developed countries the

people the civilization the culture the

habits the even the language is more

similar with the US so that includes

Canada Australia New Zealand and most of

Europe and Japan but the developing

countries their demographics young

relatively less resources and money big

interest in entertainment and social

those kinds of habits and games in

developing countries better match China

because China five or ten years ago very

much is similar to what India Indonesia

and their countries are today so Chinese

software is likely to be more successful

in these developing countries so we will

end up seeing both US and Chinese

software as quite successful but

probably to different extent in

different countries

well I don't think it's a question that

want to I don't think we ever really

have choice on technologies when a

technological tidal wave comes whether

its electricity internet or AI we have

to just accept it's coming what we can

do is prepare ourselves for the issues

and try to solve the problems that it

brings about for AI the biggest problems

people see today are privacy security

bias lack of transparency the black box

nature of AI and of course job

displacement and by AI and automation as

well as wealth inequality each of this

requires a different set of solutions on

privacy I think we need both regulations

and all kinds of regulations as well as

technological solutions for example when

electricity was became popular people

got electrocuted and circuit breakers

were invented when internet were

connected was connected to PCs viruses

spread to PCs and then antivirus

software was created so we need some

kind of technological approach to deal

with the privacy problems in the same

with security when security for example

deep fake and people hacking into AI

models so again those need to be

addressed

much like the security software or the

antivirus software in terms of bias and

the transparency we need to invest in

technologies that help AI explain itself

and also help AI developers be aware

when there might be bias created in the

software most of the bias is created as

a result of having imbalanced a training

set so if you have a training set with

95% men 5% women it might end up

discriminating or at least not

representing women in terms of the the

predictions of the model so I think the

engineers and developers need to be

trained that transparency explain

ability are an important part of

developing AI

and then tools need to be developed to

try to catch these problems before they

become really bad and this awareness and

education and then lastly on job

displacement and wealth inequality I

think several steps need to be

considered one is a shift of the wealth

because the rich is getting richer the

poor is getting poorer partly because of

technology and now it will be

exacerbated by AI so how to provide a

acceptable transfer of wealth from the

newly created super rich to help the

people whose jobs might be lost or

replaced

so that wealth transfer it needs to be

designed for each different country may

take a different approach to deal with

that problem also jobs retraining are

needed always governments should watch

for what kind of jobs are emerging as a

result of AI and what jobs are emerging

as a result of society with AI these

include not just high end jobs like data

data scientist and AI engineers but also

we're going to need a lot of robot

repair people and we're going to need

people who label data and we're going to

need services jobs because people will

live longer so they will so more people

should go into services not into doing

routine work so I think a systematic way

to examine what jobs are being created

either by AI technology or as a result

of society evolving aging and so on and

making sure that our schools and

vocational training and also government

subsidies are applied so that enough

people are moving into these jobs

finally there are social issues with

respect to some of these jobs for

example service jobs let's say a nanny

or elderly care in most countries that's

not considered the most desirable or

highest pay the job so what can be done

to entice people to go into these job

categories

does there need to be professional

companies with career paths for people

in these professions or does it need to

be a minimum wage for these jobs or does

there need to be social re-education for

people to respect these people who are

helping other people so unless the

social status and the pay is fixed it's

very hard just to say we want to retrain

people to be nannies and elderly care

and all of these things I think are

challenging and difficult but they can

be done and must be done so that the

world will embrace AI with much much

more benefits than downside

[Music]

China has been a very different market

with different language user habits so

that has been one inhibitor for European

companies to enter going forward I think

new opportunities will arise because of

the us-china tension there will be

Chinese companies that will prefer or

even be required to select European

technologies for example Huawei is

currently unable to use some American

technologies so now is a good chance for

the European substitutes to supply to

hallway so while I don't like us-china

tension but that tension does generate

opportunities for for Europe and I think

China will open its market to more to

the whole world

so whatever US was able to achieve in

helping China open this market I think

the same will apply for Europe so there

are multiple benefits

the major question I would ask is that

there are a lot of great existing large

European companies I think they will

stand to benefit I am a little worried

about European entrepreneurs if you look

at European small medium business

especially in tech as a percentage of US

and China

the numbers are disproportionately small

because if you look at giant companies

there are many great European companies

if you look at small startups that are

you know one to ten years old the

numbers in Europe are very small

compared to us in China so in order to

make this sustainable for Europe to have

a continuous set of companies and

technologies that export to China and

other countries it's important to

address the VC and entrepreneurial

ecosystem to ensure that the new ones

are emerging at an appropriate

percentage otherwise Europe has the risk

of falling