How AI is transforming our world

0:01 Despite its promise, there's plenty of concern about our AI future.

0:06 But in towns across America, one issue is already front and center.

0:11 Here's Robert Costa.

0:15 Archbald is a little corner of northeastern

0:17 Pennsylvania where coal used to be big business.

0:21 But now there is a new boom.

0:24 Data centers.

0:26 There are multiple projects happening at the same time.

0:29 So, we're not just looking at one, we're looking at many.

0:33 Yes.

0:33 Kaylee Cornell, a teacher, and Sarah Gabriel, an ICU nurse,

0:37 run a neighborhood association in this community of 7,000

0:40 where a half dozen data centers have been proposed.

0:44 What's it like living in Archbald?

0:46 It is a really nice small town.

0:49 It feels like home.

0:50 I just really love it and I can't imagine being anywhere else.

0:54 But they and others are alarmed about the possible impact on the environment,

0:58 their electric bills, and the town's character.

1:01 It's going to just completely change the landscape.

1:04 Anywhere there's trees, there's probably not going to be any longer.

1:08 We'll stop it if we can help it.

1:10 Stopping it won't be easy.

1:12 Demand is huge.

1:14 Tech companies say they need these massive structures

1:18 full of computing power to fuel the AI revolution.

1:21 So developers are rushing to build them in places like Archbald,

1:26 rich in land, water, and power.

1:29 But there is an intensifying debate over

1:32 the scope of jobs and revenue they bring.

1:35 And many communities are on edge about rising electric utility bills.

1:40 Get out of here.

1:42 We made up our minds.

1:44 Go home.

1:46 In Archbald, the push to slow things down has

1:49 upended local politics and one project has hit a roadblock.

1:54 You didn't expect to fight against AI superpowers in the global economy.

1:59 Yes.

1:59 And the development of our We don't want to be.

2:01 No.

2:01 No.

2:02 So, this wasn't part of the plan.

2:04 No, not at all.

2:04 No.

2:05 We're not against AI data centers,

2:08 but because the industry is so new and unregulated,

2:12 it is concerning that if we just keep moving forward,

2:16 we're going to get to a point where like a point of no return.

2:20 Elsewhere, data centers are already a way of life with more than 4,000

2:25 and counting in operation across the country

2:27 and many more going up around the world.

2:31 Here in Lowden County, Virginia, known as Data Center Alley,

2:35 these anonymous sci-fi looking buildings seem to be everywhere.

2:40 This building right here is north of a million square feet.

2:42 But you could visualize it.

2:43 You could park maybe not just one but two aircraft carriers.

2:47 Two aircraft carriers.

2:49 Andy Power is the president and CEO of Digital Realy,

2:53 which owns and operates hundreds of data centers globally.

2:57 Just how big is this sector?

2:59 This is hundreds and hundreds of billions of dollars as an asset class.

3:04 Power knows that these sprawling sites can seem ominous,

3:09 but he is confident they are the backbone of the emerging AI economy.

3:15 The digital infrastructure that you're standing next to essentially

3:18 could be the breakthroughs that are cure new diseases.

3:21 It'll say essentially improve quality of life.

3:24 What would you say to someone who says, "Andy,

3:27 we respect what you're doing, but we just don't want it here."

3:30 I would say, "I completely understand it.

3:32 Let me tell you why this location makes

3:35 the most sense for the data centers and this infrastructure

3:38 is going to help change the world you're living in today and for years to come."

3:43 In Washington, there are fierce critics calling for a moratorum

3:47 on data center construction until tougher

3:50 AI regulations are enacted by Congress.

3:53 I fear that Congress is totally unprepared for the magnitude

3:59 of the changes that are already taking place.

4:04 But President Trump and many Republicans say

4:06 data centers are vital for our future economy.

4:11 How important is it for America to win this race in AI?

4:16 I mean, I think it's the most important question facing our country.

4:22 Senator Dave McCormack, a Pennsylvania Republican,

4:25 has worked to bring data center investments to his home state.

4:29 I think the net benefit for Pennsylvania is enormous.

4:33 He says data centers need to work with, not against communities,

4:37 but wants companies building in his state.

4:41 There's got to be a covenant.

4:42 There's got to be a promise that goes from the community

4:45 to those that are developing in the community of what kind of job creation,

4:49 how they're going to protect the environment,

4:51 how they're going to protect water supply,

4:53 uh how they're going to protect energy costs.

4:55 When a community looks at the totality of the jobs,

4:59 the tax revenue, um the new roads, the libraries, the schools,

5:03 the the opportunity of jobs for their the their kids.

5:07 I think these are pretty compelling.

5:10 Back in Archbald, the data centers continue to move forward.

5:14 And for Kaylee Cornell and Sarah Gabriel,

5:17 it's a reality they're grappling with every day.

5:21 People live here because of the quality of life.

5:23 Yes.

5:24 Obviously, you need jobs,

5:25 but this would be intrinsically changing the character of Archable Burrow.

5:29 So, you feel like you still have a fighting chance to stop this?

5:32 Yeah, absolutely.

5:33 Oh, yeah.

5:33 I mean, we have no other alternative.

5:36 I mean, it's our home.

5:38 We have to fight.

5:40 Yes.

5:42 Daario Amade is at the center of a new kind of firestorm.

5:46 It's about the principle of standing up for what's right.

5:49 What's wrong in his view is why his AI

5:52 company Anthropic has been banned from the federal government.

5:56 It feels very punitive and inappropriate given

5:59 the amount that we've done for US national security.

6:02 Anthropic created Claude, an AI chatbot you might use at work or school.

6:07 Since last summer, its government version has been deeply

6:11 embedded in military intelligence and classified operations at the Pentagon.

6:17 This past week, in the leadup to the attack on Iran,

6:20 the Defense Department demanded anthropic hand over

6:23 its AI without restrictions for lawful military use.

6:28 The company refused.

6:30 We have these two red lines.

6:32 We've had them from day one.

6:34 We are still advocating for those red lines.

6:37 We're not going to move on those red lines.

6:40 Those red lines not allowing Anthropics

6:43 AI to perform mass surveillance of Americans and prohibit its AI from powering

6:49 fully autonomous weapons without any human involvement.

6:54 It doesn't show the judgment that a human that a human soldier would show,

6:59 friendly fire or shooting a civilian or just the wrong kind of things.

7:04 We don't want to sell something that we

7:06 don't think is reliable and we don't want

7:07 to sell something that could get our own

7:09 people killed or that could get innocent people killed.

7:12 It's a question of who should control the most advanced technology ever created,

7:16 a private tech company or the federal government.

7:21 Do you think that Anthropic knows better than the Pentagon here?

7:25 One of the things about a free market and free

7:28 enterprise is different folks can

7:31 provide different products under different principles.

7:35 Our model has a personality.

7:37 It's capable of certain things.

7:39 It's able to do certain things reliably.

7:41 It's able to not do certain things reliably.

7:43 And I think we are a good judge of what our models

7:46 can do reliably and what and what they cannot do reliably.

7:49 After several days of talks, President Trump on Friday directed the US

7:54 government to halt all use of anthropics AI,

7:57 cancelling more than $200 million in federal contracts.

8:02 Defense Secretary Pete Hegsth labeled it

8:05 a supply chain risk to national security,

8:08 a first for an American company.

8:10 President Trump has called Anthropic a left-wing woke company.

8:14 Is this decision at all driven by ideology?

8:17 I look I can't speak for what you know I can't

8:20 speak for what other parties are doing and what they're doing

8:24 but you and here in anthropic Yeah.

8:25 Look we we but we we I think have tried to be very neutral.

8:30 So this idea that we've somehow been

8:33 partisan or that we haven't been evenhanded.

8:36 We've been studiously evenhanded.

8:38 Critics call this an abuse of power.

8:40 What the Pentagon is doing and what the White House is doing.

8:43 Do you believe this is an abuse of power?

8:45 You know, again, I would return to the idea that this is unprecedented.

8:49 But is it an abuse of power?

8:51 You know, this has never happened before.

8:53 This designation has never happened before with an American company.

8:58 And I think it was made very clear in some of their statements,

9:01 in some of their language that this was retaliatory and punitive.

9:05 I don't I don't I don't know what else what else to call it.

9:08 Retaliatory and punitive.

9:10 As Amade and Anthropic face a government ban,

9:13 his main rival, Sam Alman of Open AI,

9:16 the maker of Chat GPT, struck his own deal with the Pentagon on Friday.

9:22 Amade says Anthropic plans to take legal action.

9:26 All we've seen are tweets from the president and tweets from Secretary Hexa.

9:30 And he says Anthropic remains at the negotiating table hoping to talk.

9:35 If you had a moment with the president right now, what would you say to him?

9:38 We are patriotic Americans.

9:40 We have done everything we have done has been for the sake

9:44 of this country for the sake of supporting US national security.

9:48 We believe in defeating our autocratic adversaries.

9:52 We believe in defending America.

9:53 The red lines we have drawn we drew because we we we

9:58 we believe that crossing those red lines is is contrary to American values.

10:03 Disagreeing with the government is the most American thing in the world.

10:06 And we are patriots.

10:08 In everything we have done here,

10:09 we have stood up for the values of this country.

10:19 In recent months, we've spent plenty of time

10:22 focusing on the mysteries of artificial intelligence,

10:25 but nothing quite like this.

10:27 With Luke Burbank, seeing is believing.

10:32 Last fall, if you were one of the people attending

10:35 the prestigious Summit of the Future at the United Nations,

10:38 chances are you also saw a truly unique piece of art.

10:43 Something pulsing and swirling and essentially living

10:46 the creation of Rafi anodol powered by artificial intelligence.

10:52 I think AI is inhuman but I think AI is a mirror for humanity.

10:55 So we just invent a mirror and this mirror has an intelligence reasoning.

10:59 So it will be like a very interesting journey for humanity.

11:03 That mirror as he calls it is large nature model coral which uses

11:08 an AI algorithm to draw on over

11:10 a 100 million photographs of coral to hallucinate,

11:13 you heard that right, a series of new totally original coral images.

11:18 As his work creates new digital coral,

11:20 he hopes to draw attention to the coral die off worldwide.

11:24 It will be a painful process most likely for the humanity in general

11:27 when we are all transparently in front of the mirror with AI.

11:31 If that idea that artificial intelligence can be

11:34 a living mirror sounds kind of out there,

11:36 well then welcome to the world of Rafi Anodol

11:39 where big ideas lurk beneath bigger than-l life art installations.

11:44 His work unsupervised machine hallucinations was the first generative art

11:49 ever acquired by the Museum of Modern Art in New York.

11:54 We train AI on 138,000 older artworks at Museum of Modern

11:58 Arts um archive and let AI never ever replicate existing

12:04 artwork in the archive but find a way a new vocabulary

12:08 a new language a new type of art and to do

12:12 that even we work like two years to let AI

12:15 not to show us real things but hallucinate and dream so

12:18 it's possible to use AI to literally create these dream

12:22 machines which is the opposite of what they are designed for.

12:25 The pigment

12:26 Anad Doll's mother who's from Turkey says she was shocked literally

12:30 by an exposed wire during one of his early experiments with technology.

12:37 She explains Rafi would dismantle things and then try to put them back together.

12:42 Sometimes successfully, sometimes not.

12:46 When Anadal arrived in LA from Istanbul in 2012 as a student at UCLA,

12:52 he expected the Los Angeles skyline to be something out of the film Bladeunner.

13:01 I arrived so late, 2:00 a.m.

13:04 Apparently, at that time, the buildings just shut down the lights.

13:07 I didn't like see the the the LA I was looking for.

13:10 I saw something dark.

13:11 But that was a night I believe that one day I hope I can

13:15 just give a life through this beautiful

13:17 like buildings and facades like the bladeunner.

13:21 Let the building dream remember and just hallucinate.

13:26 Fast forward a decade and Anadal's dreams are a reality.

13:30 Whole buildings are his canvas.

13:36 from the Las Vegas sphere to Casa Batllo

13:39 in Barcelona and Walt Disney Concert Hall in Los Angeles,

13:44 now his adopted hometown.

13:47 And soon, Anadal hopes to open his own museum in downtown LA called Dataland.

13:53 Dataland is a place where we interact with data,

13:56 touch, taste, learn, hear, watch data.

14:01 It's making invisible visible, inventing a new medium of storytelling,

14:05 inventing a new ways of communicating with AI,

14:08 with data, something we don't see, but we know that exists, right?

14:12 That type of like new worlds to explore and and to discover.

14:16 The space will feature AI dreams of the disappearing Amazon rainforest

14:22 and AI generated aromomas of flowers that don't actually exist yet.

14:28 We typed exactly the scientific names of the flowers

14:32 to be sure that AI understands their type, their color,

14:35 their form and shapes and then AI can connect them together

14:39 to make this AI dream of all the flowers of Amazonia.

14:43 How long is this?

14:44 Basically no end.

14:46 It can dream all these flowers many many times in different forms and shapes.

14:49 It's can be infinite.

14:51 All of this could cause one to ask the question if the only

14:55 thing dreaming bigger than artificial intelligence

14:58 just might be Rafi Anodol himself.

15:08 We've heard plenty about how artificial intelligence is reshaping the workplace,

15:13 but what could AI bring to the future of warfare?

15:17 With David Martin, we look to the skies.

15:21 When you look at it from this angle, it's a it's a pretty sleek aircraft.

15:24 Think of what goes on in this hanger at Egglund

15:27 Air Force Base in Florida as Top Gun AI.

15:31 This is one of our uh main test vehicles.

15:34 Major Trent McMullen is a fighter pilot learning to fly alongside the XQ58,

15:40 a drone piloted by artificial intelligence and a test

15:44 bed for a new kind of Air Force.

15:46 Have you flown with that?

15:48 I have.

15:48 I've flown safety chase on it for several missions,

15:51 messaging back and forth with the autonomy on board.

15:54 How well does it maneuver?

15:55 It takes a little bit of getting used to flying next to it.

15:58 As humans, we fly very smooth, but it can roll and uh and fly

16:03 a little bit snappier than maybe a human pilot would.

16:06 Sounds like a rougher ride.

16:07 It could be a little bit rough of a ride, but there's no human on board.

16:11 Artificial intelligence is on board and now it's learning how to fight.

16:15 So, what are some of the tasks that you could assign that aircraft?

16:19 Hey, there's an adversary out there.

16:21 Perform an intercept on that adversary.

16:23 So, we've been able to give it some

16:24 of the basic blocking and tackling of air combat

16:27 uh that we as human pilots also train

16:29 uh on when we're first learning how to fly.

16:33 The XQ58 blasts off like a rocket,

16:36 but a fullscale model took off from a runway for the first time in August.

16:41 Could you put weapons on it?

16:43 Absolutely.

16:44 General Adrien Spain, head of air combat command,

16:48 is drawing up plans for operating AI piloted drones alongside manned aircraft.

16:54 You told them to go out in front

16:55 and to execute an attack on a complex set of targets.

16:59 Uh, and they will do that.

17:01 Are AI drones capable of doing what you just described today?

17:06 Absolutely, we can.

17:07 Uh, it is possible today.

17:09 An AI piloted F-16 has already held its own

17:12 in a limited dog fight against an experienced fighter pilot.

17:19 Top Gun AI.

17:20 Other F-16s are being rewired for more realistic combat.

17:25 Still got a uh cockpit.

17:27 Does it still have a pilot?

17:28 It does.

17:29 They can engage the autonomy on and then remain as a safety pilot.

17:33 So once the AI goes on, the hands come off.

17:35 Yep.

17:35 They'll they'll they'll be monitoring the system

17:37 and ready to take over at a moment's notice,

17:39 but we'll also have real live aircraft out there for it to fight against.

17:43 Those jets will be piloted by real uh fighter

17:45 pods trying their best to uh outsmart the AI.

17:49 How does the AI get in the plane?

17:51 Yeah.

17:52 Is it all right if I use your notebook?

17:55 Yeah.

17:54 Uh so the pilot's actually going to hand

17:56 carry something like a cartridge about this size out

17:59 to the aircraft and then we're going to have

18:01 a receptacle for them to put that in.

18:03 So it's plug and play.

18:04 It's that simple.

18:05 Are we witnessing a revolution?

18:07 If we continue down this path, it has the potential to be a revolution.

18:12 Retired Air Force Lieutenant General Clint Heinote

18:14 says it's a revolution born of necessity.

18:17 The Air Force was so good for so long that it didn't need to change.

18:21 Now it needs to change and it's trying to figure out how.

18:25 changed because the Chinese Air Force,

18:28 which recently showcased its newest jet fighters and its own AI drone,

18:33 could be more than a match for the US Air Force.

18:36 If we have to fight China, we're likely doing it in their front yard.

18:41 And that means they can bring many,

18:44 many more things to bear than we can because it's just so far away.

18:49 you're having to achieve kill ratios of 10:1,

18:54 15 to1, 20 to1 to even stay in the game.

18:58 So, how do these war games come out when

19:00 uh American pilots are going up against 20 to1 odds?

19:04 The war games don't turn out very well.

19:07 We lose.

19:07 We lose.

19:09 The Air Force is counting on AI drones to even the odds.

19:14 What does AI bring to an aircraft that a human pilot doesn't?

19:19 The big thing with uh artificial intelligence is

19:22 its ability to handle large amounts of data.

19:26 A human out in a complex air combat environment,

19:29 there's just no way to absorb all of it.

19:33 Artificial intelligence might be able to take all of the data information

19:37 and then process that uh very quickly and then make realtime decisions.

19:42 AI drones will be about half the length

19:44 of a man jet fighter and one quarter the cost, 20 to$30 million each.

19:51 You could buy more airplanes, put them in the field,

19:55 and still not break the bank.

19:58 The key would be is that you don't have to bring that human operator home.

20:03 You actually can take more risk.

20:05 How many of these are you planning to fuel?

20:07 Uh we expect to have uh uh up to a thousand

20:11 uh aircraft eventually and uh 150 by the end of the decade.

20:15 These drones aren't just going to sit

20:18 in a hanger waiting for war with with China.

20:21 What are they going to do in peace time?

20:23 It's pretty wide open.

20:24 Could you send up AI drones to intercept those Russian

20:28 bombers that come down off the coast of Alaska?

20:31 Could you?

20:31 Yes, you could do that.

20:33 Those intercepts can turn nasty in an instant.

20:36 Last year, a Russian fighter rocked an American F-16.

20:40 So, AI drones would have to be prepared to shoot.

20:43 Is AI going to be making life or death decisions?

20:47 Absolutely not.

20:48 Absolutely not.

20:49 Uh, the human who's controlling the AI will make the life and death decisions,

20:54 at least for now.

20:56 Increasingly, militaries around the world, including the United States military,

21:00 are going to be pressured to give the machines

21:04 more leeway in making those life or death decisions.

21:08 Are we talking about giving them uh the capability to fire on their own?

21:16 The United States military is investing

21:18 in the experimentation that you would need

21:21 to be able to produce the types of platforms that could fire on their own.

21:25 if you gave them that option.

21:26 And if the other side lets AI make the decisions, what happens?

21:32 Uh I think they do so at their own peril because the AI can be fooled.

21:35 The AI can be overwhelmed.

21:38 It can give you false outcomes.

21:39 We've seen that AI can hallucinate.

21:42 Uh so it's not a guarantee of success.

21:44 What it guarantees is that it will do something quickly.

21:47 Don't you at least have to give your pilots that option

21:52 to go full autonomy and and just let it go?

21:56 I I don't think America's comfortable with that yet.

21:58 Um I'm not saying that couldn't be a a future

22:01 world that we live in where we trust it,

22:03 but I don't think we're there right now.

22:04 Certainly, and certainly not to start.

22:06 We have to build that trust over time.

22:09 At Top Gun AI, pilots keep putting in the reps.

22:12 test flight after test flight, simulation after simulation to build that trust.

22:19 Would you rather go into combat with a human wingman or an AI wingman?

22:24 When we're talking about uh the the threats of tomorrow,

22:27 if I can send an uncrrewed asset into a high-risisk environment,

22:31 I'd rather do that than send a human pilot.

22:40 We begin this Labor Day weekend with David Pogue

22:43 and a workplace topic on the minds of many these days.

22:47 Artificial intelligence.

22:50 Programs like Chat GPT use artificial intelligence

22:53 to do thinking or writing or creating for us.

22:58 Write a memo about laying off 10% of our company.

23:01 Dear team, I need to share some very difficult news.

23:04 Blah blah blah.

23:07 Summarize previous lawsuits about security deposit disputes.

23:12 And there are the lawsuits.

23:15 Mock up three box designs for our new placky toothpaste.

23:20 And about 30 seconds later, boom.

23:23 I mean, pretty amazing, but also a little terrifying.

23:26 I mean, what happens to the people who used to do those jobs?

23:30 I've applied to probably over a hundred jobs in the past.

23:35 I don't know, six months.

23:37 And yeah, I have none of them are landing.

23:42 Olivia Fair graduated four years ago.

23:44 She's had a series of short-term jobs.

23:47 One was in TV production transcribing interviews.

23:51 But now they don't have a bunch of people transcribing.

23:54 They have maybe one person overseeing all of that and AI doing the rest,

24:00 which I think is true for a lot of entry- level positions.

24:04 and it can be a really useful tool for those people doing that work,

24:07 but then there's less people needed.

24:10 This is a tough year.

24:11 Job postings um have declined year-over-year by 6.7%.

24:16 Younger job seekers, specifically those who are recent grads,

24:19 are having a harder time finding work.

24:21 Laura Olrich is the director of economic research at Indeed,

24:25 the job listings website.

24:28 The decline in jobs for recent grads seems to correlate with the rise of AI.

24:36 Yes.

24:35 Is there a cause and effect?

24:37 I think there is a cause and effect,

24:40 but it's maybe not as significant as a lot of people would think.

24:44 If you look specifically at tech jobs,

24:46 job postings are down 36% compared to prepandemic numbers.

24:50 But that decline started happening prior to AI becoming commonly used.

24:56 Wow.

24:56 So if it wasn't AI that caused this decline, what was it?

24:59 There was a real hiring boom in tech and in some other sectors,

25:03 right, in kind of the 2021 2022 time period, post-pandemic job boom.

25:08 And so quite frankly, I think some companies overhired.

25:12 The uncertain national situation doesn't help either.

25:14 Tariffs, taxes, foreign policy.

25:17 Some other people have used the analogy of like driving through fog.

25:20 If it's foggy, you slow down a bit, but if it's really foggy, you pull over.

25:24 And unfortunately, some companies have pulled over to sit

25:27 and wait to see what is going to happen.

25:29 That sounds a little more nuanced than the headlines,

25:32 which make it pretty clear that AI is coming for our jobs.

25:37 I mean, are we all just going to wind

25:38 up as sedentary idol blobs like in the movie Wall-E?

25:42 Well, then what do you want to do?

25:44 I don't know, something.

25:49 Wow.

25:50 I read today a interview where guy said, you know,

25:52 by 2027 we will be jobless, lonely, crime on the streets.

25:57 And I said, how do I take the other side of that bet?

25:59 Cuz that's just not true.

26:00 I'm sure of that.

26:01 It's the subject of every lunch conversation.

26:03 David Otter is a labor economist at MIT.

26:07 My view is look, there's great potential and great risk.

26:10 I think that it's not nearly as imminent

26:12 on either direction as most people think.

26:15 But what it does seem to do is relieve the newcomers,

26:19 the beginning incoming novices we don't need anymore.

26:23 This is really a concern and judgment expertise.

26:26 It's acquired solely.

26:27 It's acquired through immersion, right?

26:28 You know, how do I care for this patient

26:30 or land this plane or remodel this building?

26:33 And it's possible that we could strip out so much

26:36 of the supporting work that people never get the expertise.

26:40 I don't think it's an insurmountable concern,

26:42 but we shouldn't take it for granted that it will solve itself.

26:46 So, let's cut to the chase.

26:47 What are the jobs we're going to lose?

26:50 We analyze 2800 specific skills and 30% of them

26:55 could be at least partially done by AI.

26:58 So 70% of job skills are not currently at risk of AI, right?

27:03 In its current form.

27:06 So which jobs will AI be likely to take first?

27:09 Most of it is jobs sitting in front of a screen,

27:12 software coding, that's a big one.

27:14 Accounting, copywriting, translation,

27:17 customer service, parallegal work, illustration,

27:20 graphic design, songwriting, any kind of information management.

27:23 As David Otter says, what will market demand be for this thing?

27:27 How much should we order?

27:28 How much we keep on stock?

27:29 AI will have a much harder time taking jobs involving empathy,

27:33 creativity, or physicality, healthcare, teaching, social assistance,

27:39 mental health, police and fire, engineering, contracting, construction,

27:44 wind and solar, tourism, the trades like plumbing and electrical.

27:48 And don't forget about the new job categories that AI will create.

27:53 A lot of the work that we do is in things that we just didn't do,

27:57 you know, 50 or 100 years ago.

27:59 All this work in solar and wind generation,

28:02 all types of medical specialties that were unthinkable.

28:05 So, you can't sit here and tell me what the new fields and jobs will be.

28:09 No.

28:10 Right.

28:10 We're bad at predicting where new work will appear,

28:13 what skills it will need, how much of it there will be.

28:15 But based on what you've studied and the history of labor,

28:19 you can say it's likely that there will be something new.

28:22 There will be new things.

28:24 Absolutely.

28:24 So, it sounds like you don't think we are headed to becoming a nation

28:28 of people who cannot find any work

28:30 or spend the day on the couch watching Netflix.

28:32 No, I don't see that.

28:33 Of course, people will be displaced.

28:35 Certain types of occupations will disappear, right?

28:38 People will lose careers.

28:39 That's going to happen.

28:40 Okay.

28:40 But we might actually get much better at medicine.

28:43 We might figure out a way to generate

28:45 energy more cheaply and with less pollution.

28:47 We might figure out a better way to do

28:49 agriculture that isn't so land intensive and so ecologically intensive.

28:53 Whatever is going to happen will likely take a while to happen.

28:56 The latest headlines look like these.

28:58 Few signs that the tech is taking jobs.

29:01 AI hasn't taken your job.

29:03 May not steal jobs after all.

29:07 Until then, Laura Olrich has some advice for young job seekers.

29:11 The number one piece of advice I would give is move forward.

29:14 So whether that is getting another job, getting a part-time job,

29:17 finding a post-graduate internship, reach out to the professors that you had.

29:21 They have a whole network of former students, right?

29:24 Uh reach out to other alumni who graduated from the school

29:27 you went to or majored in the same thing you majored in.

29:29 It might be what gets you a job this year.

29:32 So far, Olivia Fair is doing all of the above.

29:35 So you're interested in creativity and writing and production.

29:39 So, let me hear as a human your pitch

29:43 why you'd be better than AI doing those jobs.

29:50 Okay.

29:49 Uh, I'm a person and not a robot.

29:58 The lowly grocery cart is on a roll.

30:02 Jonathan Vigliotti checks it out.

30:07 Shopping carts may help haul the food we'll feast on this Thanksgiving.

30:11 But when their job is done, they rarely get the respect they deserve.

30:17 But in some American grocery stores, shopping carts aren't just respected.

30:23 It's kind of the highlight of our week.

30:24 They're beloved.

30:26 Places like Price Chopper outside Kansas City where we met Pat and Mike Foster.

30:31 There's been a couple times we've come

30:33 that it's not been available and it it's disappointing.

30:38 Here, a new shopping cart is flying off the rack and down the aisle.

30:43 A digital screen with GPS guides people around the store.

30:48 Cameras equipped with AI technology rapidly identify

30:52 and scan products when they're placed in the basket.

30:55 Coupons are instant, and you can check out and pay right on the handlebar.

31:00 You just lift the flap here.

31:02 It's easy to just look and say the shopping cart is what it is.

31:06 Casey broker with Price Chopper began test

31:08 driving the smart carts a few months ago.

31:11 Analog and digital.

31:13 You look at the caper cart and it's

31:15 really this benefit that helps your customer know

31:18 exactly where they are and also is going

31:20 to help us from a sales perspective as well.

31:22 It almost turns a chore into something that's engaging, fun.

31:28 Everybody goes to the grocery store.

31:30 Everybody's grabbing a shopping cart.

31:31 It's a very familiar behavior.

31:33 Take me through what we're looking at here.

31:35 A behavior David Macintosh with Instacart wanted to turn into a game.

31:40 It taps into retailers existing loyalty programs.

31:42 He showed us how Caper Cart works at Instacart's San Francisco headquarters.

31:47 When I put items into the cart,

31:49 you see these four AI powered cameras pick it up.

31:52 So you see how how that

31:53 So that just scanned it as you put it in the cart.

31:55 That's it scanned it.

31:56 It turned green.

31:57 Um, and you can see a coupon popped up that I can clip.

32:00 The cart is even outfitted with a scale.

32:03 Just toss them right in.

32:04 And you can see the scale update.

32:06 That's cool.

32:07 Instacart is betting big,

32:09 spending a whopping $350 million to acquire the technology in 2021.

32:16 Shopping has come a long way in the last century.

32:20 This is where it all began for the shopping cart.

32:22 This is it.

32:24 The world's first cart was invented in 1937 by an Oklahoma City groceryer

32:30 named Silvin Goldman who put wheels and a handle on a folding chair.

32:36 He noticed that people stopped buying produce,

32:39 buying their material when their basket was full.

32:43 So he came up with this idea of putting

32:46 a basket on some sort of cart or trolley.

32:49 Steve Velasquez is a curator at the Smithsonian's

32:52 National Museum of American History in Washington, DC.

32:56 So, the shopping cart is born in the 1930s.

32:59 Does it catch on quickly?

33:00 It does.

33:01 He hired actors to pretend shop to fill the carts to walk around

33:06 and then people started using these carts and that's when it really took off.

33:11 Shopping carts have grown over the decades

33:13 as markets supersized and refrigerators kept food fresher longer.

33:19 Is it really necessary to reinvent the wheel?

33:23 Absolutely.

33:23 When you think about the shopping experience in store,

33:26 it hasn't really changed that much in the last 80, 90, 100 years.

33:30 But think about how much it's changed online.

33:33 What's happened is that online has transformed customer expectations.

33:38 The good old shopping cart, once overlooked,

33:42 once again changing the way we buy our food.

33:52 We've all heard about the promise and perils of artificial intelligence,

33:56 but lest you have any doubts about its potential,

34:00 consider this report from our Lee Cowan.

34:03 You may think that I'm talking foolish.

34:07 Even if you're not into country,

34:09 you can't help but bathe in the baritone that is Randy Travis.

34:20 Forever and ever.

34:22 Amen.

34:23 His voice goes down like whiskey.

34:25 Always has.

34:27 But on the other hand,

34:30 ever since his first album in 1986 with record producer Kyle Lenny,

34:35 I thought if we sold 40,000 copies, they might let us make a second record.

34:40 And

34:40 what did that first?

34:41 About 4 million.

34:43 I'm a genius.

34:45 There are three wooden crosses on the right side of the highway.

34:52 Randy Travis and that voice helped build

34:55 Warner Music Nashville into what it is today.

34:58 Chris Lacy is the label's co-chair and co-president.

35:02 What was it about his voice that was so unique?

35:04 How do you describe a thing that just hits you in the center of your chest,

35:09 you know, and unconsciously makes you feel full and familiar and known.

35:14 How do you describe something like that?

35:16 You know, last night I dug your picture out from my old dresser drawer.

35:21 His was a god-given talent and then fate took it all away.

35:27 Another serious medical blow to country music star Randy Travis.

35:31 He suffered a stroke last night at a hospital in Texas.

35:34 It felt like it could be the end of an era.

35:36 I thought we would we were going to lose him.

35:41 Okay.

35:40 He was given just a 2% chance of surviving.

35:44 Besides the paralysis,

35:45 the area of his brain that controls speech and language was hit the hardest.

35:51 Music is what he's made of.

35:53 Music is his heart.

35:54 It's his soul.

35:55 His wife, Mary, does most of the talking these days.

35:59 He knows what he wants to say, but he

36:01 just doesn't come out.

36:03 You've made peace with it.

36:05 Yeah.

36:07 Yep.

36:07 And how do you do that?

36:09 Oh, no.

36:11 Three years after the stroke, with Mary at his side,

36:14 Randy Travis was inducted into the Country Music

36:17 Hall of Fame and gave Amazing Grace new meaning.

36:23 Amazing Grace, how sweet.

36:30 There wasn't a dry eye in the house.

36:35 Hide me.

36:38 I told you so.

36:41 In his absence though, he's heard a lot of AI generated fakes of his voice.

36:47 A simple Google search brings up sites like these promising

36:50 to convert anyone's vocal into a Randy Travis sound alike.

36:55 It doesn't sound good and it doesn't sound real because it's not.

36:59 And that got his record label thinking.

37:02 We started with this concept of what would AI for good look like for us?

37:07 And the first thing that came to mind

37:08 was we would give Randy Travis his voice back.

37:11 There were, however, more than a few skeptics.

37:14 It sounded like, you know, like a parlor trick or something.

37:20 But Kyle Lenning with Ry's blessing decided to try.

37:23 He and Warner Music started by pulling 42 Randy Travis tracks from the vault.

37:28 operator, please connect me

37:32 like this one, the original recording of his hit 1982.

37:37 And then they stripped away the music, leaving only the vocal.

37:47 Wow.

37:51 That was half the recipe.

37:52 The other half needed to be provided by a donated or surrogate voice.

37:58 So I have to get a level on you.

38:01 In this case, that voice came from country music singer James Dup Prey.

38:05 Being a part of new Randy Travis music, I mean that's like what

38:16 that's what was put into the model.

38:18 The model, it's the secret sauce.

38:21 Here's the vocal loaded in

38:25 an AI program that takes Randy Travis's voice

38:27 and overlays it on top of James Dup Praise singing.

38:31 It's hardly an exact science.

38:34 It's not about how it sounds, it's about how it feels.

38:37 And that's not something a computer can figure out.

38:39 Not yet.

38:41 Kyle knows Ry's voice almost better than Randy does.

38:44 They worked together for 40 years.

38:46 This time though, the task was to take

38:48 a computerenerated voice and give it Ry's country heart.

38:53 Him being here and him being able to be, you know,

38:57 a vital part of the decision-making process makes all the difference to me.

39:01 In the end, they finally landed on something they felt worthy

39:05 of a voice no one but those who knew him best could replicate.

39:10 It's Randy Travis.

39:11 Ry's on the other side of the microphone.

39:13 It's still his vocal.

39:14 There's no reason he shouldn't be able to make music.

39:17 And to deprive him of that, if he still wants to do that,

39:21 that's unconscionable to me.

39:23 Two months ago, Warner Music gathered a small

39:26 circle of fellow musicians into a recording studio.

39:29 Randy sat with a cheshure cat's grin and then they hit play.

39:37 Holy crap.

39:39 The reaction to his first song in more than a decade was a mix of joy and wonder

39:52 for his wife tears.

39:55 You forget how much you missed it until you hear it again.

40:00 For Ry's country friends like Grammy winning superstar Carrie Underwood,

40:05 there was confusion.

40:07 How how how

40:10 for Cole Swindell who just won three ACM awards myself big

40:18 reminded him of why he became a country singer in the first place

40:22 for y'all letting me hear it that means a lot damn I'm glad to hear you sing

40:28 and veteran country star Clay

40:32 Walker he was simply over the Even Ry's own family hadn't heard it

40:44 until two weeks ago.

40:46 It's so weird to kind of explain like everything

40:49 that goes through your head when you're listening to it.

40:52 It was perfection

40:54 with a caveat.

40:56 We don't exactly know how to get here again.

41:00 You want to laugh at me as I try to sing this chorus?

41:03 She set her diamond ring next to another round of burden.

41:10 The second song they're working on is proving a harder nut to crack.

41:15 Sometimes angels spill whis.

41:20 Wow.

41:21 That is so crazy.

41:22 Hear how all the verd doesn't work right.

41:24 Yeah.

41:24 Yeah.

41:26 Yeah.

41:25 There's still work to be done clearly and a lot

41:28 of questions to answer about what this all means going forward.

41:34 But rebel in this.

41:35 After a decadel long absence, Randy Travis is back on the radio.

41:40 His Poststroke debut song,

41:42 Where That That Came From, was released this past week.

41:47 For Randy Travis, it's not just a single, it's a victory.

41:52 It's a life inspiration.

41:54 Speak kindly, love fully, live completely, and live leave the rest to God.

42:01 You know, sounds like there's a song in there somewhere.

42:05 We can only hope.

42:11 Where?

42:17 You say you couldn't make it to Paris to take in the latest Van Go show?

42:22 Not to worry.

42:23 Seth Don can tell us all about it.

42:31 With bold colors and dramatic brush strokes, Vincent Van Gogh,

42:36 that tortured genius, has traditionally expressed himself through his art.

42:45 But now there's also his avatar.

42:48 Van Go, what is your greatest inspiration?

42:53 My greatest inspiration, without a doubt, is nature.

42:55 He's a tad monotone, but is equipped with a wealth of knowledge.

43:00 It is in nature that I find comfort.

43:02 Artificial intelligence combs select biographies and the artist's own writings,

43:07 900 letters covering art, relationships, and his struggles with mental health.

43:14 As you can see, the interface is quite simple.

43:16 It's a digital screen and a microphone.

43:19 But behind this, there is months of research in artificial intelligence.

43:23 Anes Abastado oversees digital development for the Muse

43:27 Dorsce in Paris as this storied

43:30 museum in an old train station finds new ways to engage visitors.

43:38 Its recent Van Go exhibition included a virtual reality

43:42 experience based on the artist's pallet and that avatar

43:46 Bonjour Vincent which allows visitors to pose questions

43:50 to the artist more than a century after his death.

43:55 How were the last few months of your life?

43:58 Last months of my existence were a blend of both hope and despair.

44:01 A tumultuous combination that consumed me wholly in the end.

44:05 It was a time of great turmoil which eventually led to my premature demise.

44:10 It's remarkably personal.

44:12 Yeah, it is personal because the technology really

44:15 want to uh recreate the personality of someone.

44:19 Van Gogh was a troubled artist who killed himself at age 37,

44:24 just 10 years into his painting career.

44:27 But he was prolific at the end.

44:30 In his last 70 days of life, he was producing more than a painting a day.

44:34 and dozens of drawings, many of which were gathered at the museum.

44:39 You've read those 900 letters of Van Goes.

44:43 I asked the Van Go avatar what his greatest inspiration was.

44:46 He said, "Nature." Was the avatar right?

44:49 Yes, totally.

44:50 Totally.

44:51 Emanuel Kokari is the exhibition curator.

44:54 He notes this period of Van Go's life is marked by experimentation.

44:59 clouds painted in reverse blue brush strokes over a white background.

45:04 It's very lively because you still uh see uh the gesture

45:09 of the painter and you see the the flower here very thick.

45:12 When you're this close to the painting, you can actually see the paint.

45:15 Exactly.

45:15 We always say it's a painter of color, but above all a painter of thickness.

45:21 The Dutch artist had moved to the countryside near Paris in May

45:25 1890 after famously cutting off his ear and spending time in an asylum.

45:32 He'd come to see Dr.

45:33 Paul Gash whom he painted.

45:36 Dr.

45:36 Gash told him to forget about all your trouble problems.

45:41 Just work, work, paint, paint.

45:42 So all day long he painted.

45:45 He followed doctor's orders painting the roots

45:48 of an acacia tree at the very end.

45:51 He finished it on the day uh he he died.

45:56 And of course the symbolism is really heavy.

45:59 He wants to marry to find a house

46:03 and to root himself in family life with children.

46:07 He didn't succeed.

46:08 And he says in his last letter that was found in his pocket,

46:12 only our painting can speak.

46:14 But now those letters are speaking through a higher tech form of expression.

46:19 But even though Dr.

46:20 Gashi gave me his word,

46:22 I could not bring myself to believe that my troubled mind could ever be healed.

46:25 Some visitors could say this is a gimmick

46:28 or almost a distraction from from Van Go's art.

46:33 It's another way to enter what we know and all the knowledge about Vanog.

46:38 And what I think interesting about this interface is

46:40 that it's not us museum who choose a theme.

46:43 It's the visitor who really has what he want to know about Vongo.

46:49 We asked AI Van Go about the initial public reception to his work.

46:54 It brings me great grief that so

46:56 many of my creations remain uncknowledged and unsold.

46:59 So he recognized that his art wasn't that wellreceived.

47:04 Artificial intelligence only hints at the poetry of Van Go's writing.

47:09 He was prolific in life and as it turns out prophetic before death.

47:15 Though I cannot claim to foresee what is to come,

47:17 I hold firmly to my conviction that art,

47:19 true art, possesses the ability to transcend time

47:21 and stir the hearts of generations to come.

47:24 To me, the worth of my art lies in its power to reverberate

47:27 with others and to reveal a glimpse

47:28 into the splendor and intricacy of life itself.

47:38 You dial the number, navigate the recorded instructions,

47:42 and if you're lucky, there's a person,

47:45 a real person on the other end of the line, if you're lucky.

47:50 But more often than not, you don't get the action you'd hoped for.

47:54 So, is your call important to them?

47:57 Really?

47:58 David Pogue looks at customer service, its frustrations, and its future.

48:06 A call may be recorded or monitored.

48:10 Want to know how I spent my Tuesday?

48:12 Trying to report some fraudulent credit card charges.

48:14 Think I'm calling Global Commercial Card.

48:16 Hi.

48:17 Yes, I'm trying to report some fake charges on my credit card.

48:20 I'll connect you with our fraud team.

48:22 I see.

48:23 Thanks for being Okay.

48:26 Customer service is a real problem in this country.

48:29 Just ask literally anyone.

48:32 So, I've got a tale for your customer service story.

48:36 That said, we're sorry.

48:37 We don't have anybody right now to answer your call.

48:39 Please call back.

48:40 AT&T had me on the phone for 6 hours.

48:44 Airline gave me the runaround via email.

48:46 It kept saying it couldn't find our flight.

48:48 said, "Lady, the technician is standing 2 feet in front

48:52 of me and he has the unit in his hand." So, why is customer service so bad?

49:00 First of all, labor shortage.

49:02 Especially since the pandemic,

49:03 it's hard to find people who want the call center jobs.

49:06 The average call center worker quits after about 18 months.

49:10 Second, consolidation.

49:12 All those corporate mergers of airlines

49:14 and banks and telecom companies and cable companies.

49:17 Now there's not as much competition,

49:18 so there's less incentive to fix the problems.

49:21 So what happens?

49:22 You get a lot of unhappy customers or worse.

49:27 Needless to say, I did not stay with Verizon Wireless.

49:30 So I filed a complaint with the public service commission.

49:33 Now I'm the proud owner of a new website

49:36 where I plan to detail the rotten customer service.

49:39 And so to this day, I have upstairs in my guest room a 4K Direct TV box

49:44 that that drops out every four minutes and it's hooked up to a TV that isn't 4K.

49:51 This isn't about money.

49:53 It's about human decency.

49:57 Well, if it's any consolation, there is a group of people who may be even

50:00 more frustrated by the customer service situation than you are,

50:04 and that is the customer service agents.

50:08 When they call in, they see you as part

50:10 of the organization that's imposing the issue rather than a human being.

50:14 Believe it or not, I've had somebody curse me out completely.

50:18 They can't see you.

50:19 They're just over the phone.

50:21 So, they take advantage of that.

50:23 Of course, they call with their frustrations.

50:26 Uh they don't understand the procedures.

50:28 They have been told no multiple times.

50:31 And sometimes that brushes off on me.

50:33 I'm frustrated, too.

50:34 Um I want to help you out.

50:36 I stand up.

50:36 I'm like, I feel like I'm going to get in a state where I'm sweating.

50:39 Would you ever say to a caller,

50:41 "I need to put you on hold for a second." When you really don't,

50:44 but you just need a break.

50:46 Of course.

50:46 Of course.

50:48 Dominique Raymond, Kayla Zuniga, Janerys Ortiz, and Calvin Echavaria.

50:53 Hypoes, if you could just please provide me with your date of birth.

50:56 Work in customer support for health

50:57 companies and patient advocacy nonprofits in Orlando.

51:01 Well, right now, I don't even see that you're registered,

51:03 so there's nothing to resubmit.

51:05 I've read that turnover in call center work

51:08 is more than double the average of other jobs.

51:12 I think it's hard.

51:14 They're not necessarily looking for a resolution.

51:17 They're looking to vent.

51:20 So yeah, that can be kind of difficult and taxing.

51:23 Okay, give me just a second.

51:24 Let me Customer service is not easy.

51:27 It's not meant for everyone.

51:29 My pleasure.

51:31 So if the job can be so hard, why do they do it?

51:34 There's also a lot of redeeming qualities.

51:36 It's very mentally stimulating because every day

51:39 is a new challenge or new experience.

51:42 Beauty of it is honestly to me is to help to help out there.

51:46 It it brings me joy.

51:47 Sometimes I cry with the patient.

51:48 I'm so excited to actually help someone out.

51:52 No cable company customer call center employee has ever cried with me.

51:59 Thank you for holding.

52:00 And so for years, that's how things have stood.

52:03 Two frustrated armies of people on opposite ends of the line,

52:06 each frustrated by a dysfunctional, underfunded system.

52:10 And then this happened.

52:14 My name is Grace.

52:15 I'm a virtual front desk agent.

52:17 What can I help you with today?

52:19 Um, I I left a bag at the gate.

52:24 Can you describe the item?

52:26 It's a 9 foot Steinway grand piano.

52:29 I have good news for you.

52:31 We were able to locate an item matching that description.

52:34 Welcome to the new age of artificial customer service.

52:38 Artificial intelligence, that is.

52:40 You can adjust these personality sliders.

52:43 How much empathy should the bot show?

52:45 How much should it respond to your emotion?

52:48 Nico Bonitez is a co-founder of Grid Space,

52:51 creators of an AI phone rep called Grace.

52:54 My name is Grace.

52:56 At this moment, more than 100 hospital chains, airlines,

52:59 phone companies, and banks use Grace to take customer calls.

53:04 It doesn't sound like you're talking to a robot.

53:06 You hear the ums and uh all the imperfections

53:09 that uh that indicate that someone is listening.

53:12 Anthony Scaderi is another Grid Space co-founder.

53:16 And so if you have a system that feels like

53:17 it's listening to you and doesn't make you restate yourself,

53:20 it's much less frustrating as a user interface.

53:22 How does grace make life easier for the humans at the call center?

53:27 Well, that's a really hard job and it's,

53:29 you know, it's just people yelling at you all day.

53:31 And so, one way grace can help is by being

53:33 kind of a bit of a blast shield by, you know,

53:35 first someone calls in, they state their intent,

53:38 they complain, the bot tries their best to help them,

53:40 and then the person can kind of come in as the hero.

53:42 I mean, that's in general how automation really helps people,

53:45 which is it enables people to do the stuff that people are best at.

53:49 But then for the other 90% of calls

53:51 where people are calling in with password reset,

53:53 password reset, password reset, or whatever, uh, Grace can just deal with that.

53:57 Grid Space set up a fake airline 800 number so I could witness

54:01 firsthand how Grace hands you off

54:03 to a human when the problem requires creativity.

54:07 Thanks for calling CPS airline.

54:09 Um, I have a service animal.

54:11 Can I bring my service animal?

54:13 Yes, you are allowed um to bring your service animal on board.

54:18 Okay.

54:19 It it is a tiger.

54:21 Is that okay?

54:22 I'm sorry, but I um don't have any

54:25 information about bringing a tiger as a service animal.

54:27 I'd be happy to transfer you to an agent.

54:30 So, common sense is not her strong suit.

54:32 If there's no policy about tigers, she's not going to invent one.

54:36 Unrestrained AI acting as a representative of your company

54:40 can get you in a lot of trouble.

54:42 And so we would rather heir on the side of being real by the book, real literal.

54:51 Meanwhile, Grace offers another benefit to us puny humans,

54:56 much less sitting on hold.

54:59 She's answering millions of phone calls simultaneously.

55:02 Yeah.

55:02 Picks up right away, even if it's a machine.

55:05 A fascinating concept, but not everyone is sure that AI is the answer.

55:10 It seems like you lose that connection,

55:14 that ability to make that good first impression

55:18 and to connect with the patient right off the bat.

55:21 It seems impersonal.

55:23 I I love the fact that we can actually be weird.

55:25 They they're live agents.

55:26 You can't do that with a AI.

55:28 Take that, Grace.

55:30 Plenty of companies are working on AI customer support systems like Grace.

55:35 But until they become widespread and reliable,

55:38 most of the people answering your calls will still be people.

55:43 We get them at their worst,

55:45 but then also at times when you can pivot and recover that situation,

55:51 then get them at their best and their most grateful.

55:54 They'll write your name down and say, "Hey,

55:56 like I want to let your supervisor know that you helped me.

55:59 And even though I started yelling at you from the beginning,

56:02 it was not your fault.

56:04 You assisted me and you never hung up on me.

56:06 So you forgive them their anger.

56:09 Yeah.

56:09 So no need to forgive.

56:10 There's no forgiving.

56:11 I don't take that personal.

56:12 You just needed somebody to listen to.

56:15 You as well.

56:15 Happy holidays.

56:18 Byebye.