AI Is Changing Your Job—Now What?

[upbeat music]

Welcome, everybody, and thank you

for being a WIRED subscriber.

Your support makes our journalism

and everything we do here possible.

And we're so excited to be here with you

to talk about our latest issue on AI and jobs.

It is the topic on everyone's minds like,

are white collar jobs going extinct?

Employers are talking all the time about going AI native,

being agentic as they like to say.

While many employees are rightfully worrying

if artificial intelligence is gonna make their jobs extinct.

So what does this mean for all of us?

Like should we all retrain, emerge as a nation of plumbers?

We're gonna get into all of this.

Consider the options.

I'm Sandra Upson, features editor at WIRED.

And with me today are two glorious panelists.

They're so fun.

We've got Reece Rogers, WIRED software editor.

Hey, everyone.

Hi, Reece.

Thank you for coming.

He explains crucial topics for WIRED readers

about how they can get the most outta their technology.

Instead of being used by your technology,

he puts you in charge.

Kate Knibbs is senior writer at WIRED

who has covered a number of really important topics.

Hi, Kate.

And right now she's really deep into prediction markets,

the future of media, and how AI is changing the internet.

She's also the person behind Wired's brand new book club,

which we'd love to have you join,

but we'll talk more about that at the end.

Reece, I'm gonna kick to you first.

You just published a story

about something that I think touches

on a few different sides of our topic today.

It's so, it's like a quirky weird little thing.

Headline is, I spent a week recording myself

doing chores for money, who's the robot now?

There is so much packed into that headline,

but the premise was you started this like quirky new gig

that has actually been like created by the AI boom.

It's this like kind of twisted opportunity.

Tell us about it.

Why does anyone wanna watch you do chores?

Are you like really good at chores?

I would ask my partner, he might say no.

But this was such a fun one to report out.

DoorDash's Tasks app that was launched earlier this year,

really put this kind of style of gig work on my radar.

So essentially, what I did for this piece

is I strapped an iPhone to my head right here.

And then I recorded tasks around the house,

whether that's scrubbing the dishes, vacuuming,

or even just as simple as like pouring water back and forth.

And then I recorded those, and uploaded those videos

to companies that are purchasing this kind of data

to train robots.

And they call this egocentric video data,

but essentially, you can just think of it

as first-person perspective video data.

And one of the reasons

that this is kind of a rising gig work

that is currently big in countries like India,

but companies are looking to expand

into the states even more,

is that in order to train a robot to be really good,

feeding it thousands and thousands of videos

of maybe just like, let's go with my hands pouring water.

If you upload thousands of those videos,

you can use that and annotate the content, the clips,

and put it into an AI model that is improving

how these robots can work in the home.

And a lot of it is pretty far off.

But during the rise of the chatbot era

kind of a few years ago, many companies made lots of money.

I'm talking billions by annotating text data,

and sharing that, and selling that to the big companies.

So there's kind of this rush right now to be that version

to get rich off of selling this kind of data.

And currently, a lot of it is focused on chores.

So they're kind of working on humanoid kind of style robots

that can be great in the household.

But thinking about the future of work,

even beyond this just being a gig work

you might do on the side

in addition to delivering food, this is kind of bleeding

into many people's nine-to-five workplaces as well.

Right now, they're kind of experimenting

with putting these similar kind of cameras on their heads

or on their chest of people working in restaurants,

working in warehouses, and tracking their data

to kind of maybe one day try to build a robot

to replace them.

And so I feel like we're really just entering this era

of whether it's gig work, whether it's office work,

there's this increased level of surveillance

that is trying to build, improve these AI models

to replace workers.

So I think there's a lot going on here.

It's kind of icky, it's kind of fascinating,

but it's definitely the future of work.

I have to ask you how you affixed the iPhone to your head?

Yes.

Like is it a duct tape situation or?

Please tell us.

It felt like that, it felt like that.

There's actually tools that are built specifically

for capturing this kind of data.

It's kind of like the skeleton of a hat

and it has a big claw on the top

that like clamps down on your phone.

But what interesting thing is there's also startups working

on like a hat, just a straight up hat

that has cameras pre-installed.

So it's definitely something that people are working on.

That is such a weird image, though.

Like all these people with cameras strapped,

like they call it egocentric, but it's sort of the opposite.

It feels really dehumanizing.

Like I can't even like wrap my mind around just this image

of you and all these other people

just like strapping cameras being like, yes,

I actually don't matter, but what my hands are doing matters

for like my eventual replacement.

This just seems so dystopian.

Yes, it left me feeling very robotic.

Like I felt like the robot by the end of it.

And I mean it kind of is a beacon of how many jobs

are kind of being more dehumanizing in this AI era,

where you're starting to feel more like a,

you're just feeling like I am a producer of data.

Like I'm not a person with soft skills and emotions.

Like I'm just fed into this kind of machine.

But I think that also kind of makes me think another way

that like all of these workplaces

are kind of becoming more surveilled

is there's a really interesting article about Meta employees

who are pushing back recently

about how all their mouse movements are tracked

to train, you know, some kind of AI model

that could use the software similar to how they're using it.

So it goes, it's much deeper than just training robots.

Like every aspect of the future of work

is likely going to be recorded, made into data,

sold off into god knows where.

Yeah, the Meta thing is crazy.

Like the keystroke logging, mouse tracking.

It used to be that those tools were kind of like

to keep employees on track, right?

To like surveillance for the sake of like capitalism.

And now it's surveillance to replace you,

to like literally, like they're so upfront about it.

We're training AIs to do this specific work,

and it feels like such a loss of employee kind of rights,

and standing that like our ability

to, you know, have a comfortable and respectful environment

is just like going down the toilet

in exchange for AI training data.

Did you, Reece, emerge from with any broader takeaways

about the industry, about how AI is developing?

Like what was your sort of takeaway feeling?

Yeah, I think my takeaway feeling

is that this specific egocentric data collection

is going to be big for the next few years.

So whether or not this becomes the kind of gig work

that people turn to if they're laid off,

I think it could be, because whether or not

this is something that they want to do,

people might be put in a position

where they're having to train the robot tonight

to eat tomorrow.

So it's fairly dystopian.

One thing I weirdly feel optimistic about

is that in 10, 15 years I do think these robots

are going to be better.

Like I think you might not be living with one,

but it is impressive like how,

there's a lot of hype behind it, yes.

But there have been some interesting advances in robotics

over the past few years,

and I'm fascinated to see where that goes.

Though, I have like a big sense of dread

about whether or not they're actually going to be able

to achieve something that can replace a factory worker,

a chef, like whether they can, whether they can't,

I don't know if it's good if they can.

Now, I'm imagining a world where I go to a restaurant,

and my robot server has like hand gestures,

and I'm like, that kind of reminds me of a Reece actually.

He had that one tick.

[all laughing]

One of the founders I talked to for this piece,

he was completely sure that every chef in the world

will be replaced by robots in the future.

I don't think I'm that sure as he is, though.

I actually like hearing that

because it's a reminder that these CEOs don't know it all.

Like they don't necessarily understand

how much artistry and human ingenuity

is behind so many jobs.

And it's just like you can't just write that off

as something that can be automated

through mass intelligence gathering.

So I don't know, I kind of take a weird

sort of like twisted bit of hope out of that.

Last question on this story, the big one,

like how was your payday?

Did you make bank?

My payday.

Okay, so for context, my rent in San Francisco

that I split with my partner is $2,500 a month.

So like that's the scale we're working with here.

I made $21 from a,

I could have hit the gig work even harder.

You know, if I had approached this as a full time,

you know, eight hours a day,

I probably could have made a little over a hundred.

But I submitted over 800 clips and barely made over $20.

And a lot of that was relying on the one platform

that paid the best and is like much more US centric.

So many of these are offering like $10 an hour

for recorded clips.

So like it's not very good money right now.

And so, the money's just, they're not paying out very well.

But one interesting tidbit with the payout,

a Waffle Video who paid $25 an hour for clips,

they actually offer this kind of syndication style.

So if any other companies license or purchase the data

in the future for the next five years,

I'll actually make money.

So in addition to that 20, maybe another two or three bucks

will be coming my way in a few years.

Cool.

You should hold them to that.

I wanna see your like hand royalties, right?

Like gotta diversify those income streams,

especially in this day and age.

Kate, we're gonna switch to your story

that you just published.

AI is taking over the most cursed job in the world,

that was the headline.

You did this really interesting thing

that, you know, you kind of notice that we tend to overlook

that not all jobs are actually desirable,

maybe some of them can go away.

And the population you focused on was debt collectors.

How did it, yeah, how did you kind of come to focus

on this group in particular?

Yeah, so this story was born from a group chat that I'm in

with a bunch of mostly former journalists at this point.

One of my friends now works as a software engineer

for like one of the big tech companies.

And he sent us this screenshot of a recruitment email

he received from one of these AI debt collection startups.

Basically just saying, I didn't know this existed.

This seems really kind of despicable,

I want nothing to do with this.

But I looked at the screenshot,

and I was like, I wanna know everything about this.

So I went down a rabbit hole googling the company.

Quickly realized that it was one of dozens of startups

specifically focused on creating AI agents

to replace human debt collectors.

Six startups alone incubated by Y Combinator

in the past few years, like in this area.

So many, that's like such a specific little niche.

Do we really need six, you know, chances at bat

for a debt collection?

Okay.

We'll see.

And so I reached out to a bunch of these companies.

It was interesting too,

a lot of them had very similar websites.

It seemed like they'd all sort of like looked

at the same web design format and went with it,

because I ended up, none of them had emails

to have press contact them, but I was able to request demos.

And so I went through this series

of AI debt collection demos,

and learned a lot about the industry,

and learned that they are already operating

at a significant scale.

Like major financial institutions, major auto companies

are already using these startup services.

They're making millions upon millions

of phone calls per month, each startup.

So it could be tens of millions.

I couldn't get like my full hands

around like the size of the industry.

And it was kind of amazing to see

just how swiftly this had become,

like they didn't completely eradicate human debt collectors,

but this is a pretty normal thing

in the debt collection industry already,

to have bots making the outreach calls, and texts,

and sometimes WhatsApp, and Facebook market messages.

Like they do a lot of different digital outreach.

And they're quite sophisticated.

Some of the bots are able to detect

when somebody's Spanish accent

is from Mexico or Colombia for example, and adjust,

because there's research that indicates

that you respond better

to people who have a similar accent to you.

Some of them are able to,

they like create psychographic profiles

of the customers and sort of tailor the way

that they approach conversations

based on whether they think the customer

will be hostile or friendly.

It's like a surprisingly mature industry

for how quickly it sprang up.

So that was just fascinating,

and I was like, I gotta write about this.

For a while though, I wasn't sure if I was gonna pull it off

because it was super easy to get all of these startup pros

on the phone, much harder to find humans

who had been contacted by these debt collectors

who were willing to talk on the record.

I did have a breakthrough, though.

I ended up finding Ben,

who's sort of the main subject of the story,

on a forum dedicated to the TrueAnon podcast

where he had posted an account of his interaction

with one of the chatbots.

And he sort of broke the story open for me.

And I'm really glad it's out in the world.

Thanks, TrueAnon, you've done us good.

But there were some interesting things,

'cause you talked about how sophisticated this is,

and how it's actually rolled out quite widely

for something as young as this,

but it wasn't like it was a seamless experience.

It wasn't like these chatbots were just killing it.

No, they kind of suck, actually.

So maybe, like, the backend looks perhaps more sophisticated

than the front end executes what's available on the backend.

Like the bot that we call,

we were able to call one of the bots back,

and so I was able to see how they interacted with customers,

and it was just pretty glitchy.

And eventually, we put a ChatGPT audio version on

with the bot, and sort of had the bots battle each other,

but the AI debt collector just kept hanging up.

So it wasn't like it was super suave,

and convinced us to pay the debt, which was good,

'cause the debt had actually already been settled.

But yeah, they were noticeably glitchy.

This is such a new area that there's not a ton

of definitive science

on how effective these bots are compared to humans.

The companies say that they are,

but then I talked to an academic at Yale

who had done some research, and he's very skeptical

that they're all that effective.

He thinks they're probably worse than humans,

but companies will still choose to use them,

because they're so much cheaper.

The fact that they're like 75% cheaper.

So the fact that they're 50% shittier

doesn't really matter in the end.

So we're talking about

this very specific population of jobs

that were affected by bots.

What kind of takeaways did you have for the industry, right?

Like we're automating all kinds of things,

like replacing humans with systems in phone call situations

is like, you know, now a tale as old as time.

We've been doing this for decades

and people have always hated it.

The hate doesn't seem to really change anything, right?

Like we still are stuck talking to some form of automation

on the phone all the time.

What did this sort of suggest to you going forward

that like it's just gonna push ahead

regardless of how good it is?

Was that kind of the takeaway or how do you kind of,

yeah, square it with what's happening?

It's definitely one of the takeaways.

Like unfortunately, it's not like customer service

as we're talking about.

Like it's not like customer service telemarketer phone calls

have ever been fun.

I think these bots are basically going to become the norm.

I think it's gonna be incredibly frustrating.

We're gonna be yelling, let me talk to a human,

like more frequently than we already are.

Reporting this just really cemented in my mind

how unevenly distributed the AI job takeover is going to be.

You know, I think that there have been elements of concern

about what AI is going to do to jobs that are overstated.

There's industries that really won't be impacted as much

as we fear that they will be just because,

as you were saying, like CEOs assuming that chefs

can all be replaced by robots

is an insane and incorrect take.

But there are these like certain tranches of jobs

that will be automated away,

have already been automated away,

and we're not even noticing it.

And I think that may have some unexpected side effects.

I was like, well, who really cares

if AI takes debt collector's job,

like it's a heinous industry,

we should abolish it altogether.

So like if, you know, someone's gonna lose their job,

I'm not gonna weep profoundly for the AI debt collector,

but I don't know, maybe I should be weeping

for the AI debt collector.

It could, you know, someone who might have trained

as an AI debt collector and then gone on to greater things,

no longer has this entry-level training ground

in which to play.

Yeah, like what's the fallback, right?

So you were a debt collector, what are you gonna do now?

Like is it back-

I guess wear one of those hats.

Could be worse.

Yeah, wear a hat.

We'll send you a hat.

I'm gonna open up to reader questions.

Thank you to those who submitted questions

and stories in advance.

You can still add your own

in the comment section right below this livestream player,

like on this same page.

So I think we're gonna talk about, let's see,

we've got reader named Skynet.

Skynet asks us, As a writer and creative person in general,

I am deeply conflicted about when and how to use AI.

How can I balance engaging with AI

in ways that are productive to my work

while also preserving the integrity of my craft.

I think this gets at the heart of like a lot

of what's in people's minds today.

Kate, Reece, I feel like you both

have really interesting perspectives on this.

Who wants to go first?

I wanna hear from both of you.

Take it away, Kate.

I will.

I'm actually writing, I just handed in a story

that should be going up soon on www.wired.com.

Check it out.

That grapples with this question

because it's a story, not to spoil it,

but it's a story that looks at where we are drawing lines

around ethical use of AI in writing.

And I don't have like a very, you know, neat and tidy answer

for you, because this is really ultimately

going to come down to each person's assessment

of their own ethics and morals.

Some people in the writing space

feel that there is no ethical way to use AI

because so many of the most popular AI tools

were trained on copyrighted materials

without people's permission.

Others have more, you know, nuanced or hypocritical takes,

take your pick.

Like, I get that argument truly but I do use AI

in my own, you know, writing practice in a way.

Like I use an AI transcription tool.

And you know, sometimes I'll query ChatGPT or Claude.

I also use Claude to come up with my Freedom

of Information Act Request draft sometimes.

So those are my own personal boundaries.

I don't use it at all in reporting,

we would get fired if we did.

But everyone has to sort of come up with their own criterion

for what is acceptable to you, what is too far.

And sometimes it might come down to, you know,

how the individual companies treat workers,

how they trained their tools,

what, you know, impact those tools have.

There's so much to consider that it is truly impossible

for anyone to tell you how to use AI.

Look within your heart, basically.

[Sandra] Go deep.

Reece, yeah, how have you thought about this?

I'm sure it's come up in your life a whole bunch.

Yeah, it's such a great question.

I think thinking about if you're a writer

who, I'm writing a novel, I'm writing essays,

I feel like it's in your best interest to not denounce AI,

but to really lean into the humanness of what you're doing.

So like using AI for as little as you can, I think,

or like just not using it.

And like publicly doing that,

I think is maybe your best move as like a creative writer,

especially in the next few years.

There's so much AI slop, right?

And kind of putting the value on you as like a boutique,

I'm a skilled laborer who's great at writing,

and the AI bots just aren't as good as me.

I think that's a really good way to go as a creative person.

Kind of thinking about journalists specifically,

so moving from like novel writing to like reporting.

There was another question in that,

was like how are journalists using it?

And there was a really good piece by our AI,

our human reporter who writes about AI,

about how more tech reporters are kind of leaning into tools

like Claude to write their first draft.

And that's something,

not only does it go against our policies,

I just would never do, because I got into this job

to be a writer and I like that aspect in it's core

to what I think of our labor.

But to go about like through our AI policy,

like wired specific, what I am and am not,

like I'm not allowed to do,

like I can't generate a story, no.

But there are three things I think

that we are allowed to do that are kind of interesting.

We're allowed to use AI to suggest headline ideas

or texts for social media posts.

This is one I've actually done a decent amount

over the past few years.

Especially headlines, I'll kind of talk,

I'll just turn on ChatGPT so it's recording my audio

and just blab at it for a couple minutes.

Tell it like kind of what I'm working on,

what I want the head to be, and ask for like 20 options.

And then kind of just iterate on those options,

and like find one or two that I think are strong

to bring to my editor.

Does it work great?

Sometimes it's okay.

Sometimes the answers are like very bad.

So I feel like it's very, it's either or on that.

We're also allowed to use AI to generate story ideas.

Here's one where like it just doesn't work very well

for like WIRED.

We want a scoop, we want to be the first people

reporting on a story, we want these original ideas.

And a lot of the times, if I just ask like,

I write about software,

so I'm like what's a hot software thing

that's buzzy I should let like write about,

it's not gonna bring me the fresh stuff I need

for you, WIRED readers.

And also, we're allowed to use AI as research tool.

And this is something I've done a decent amount of too.

Kate mentioned using it to file a FOIA request,

which is like a long technical kind of email you send

to get records, and it has been really helpful for that.

Also, if I am slammed for time,

and I need a niche expert who's knows everything

about AI music apps right now, today,

like I might ask ChatGPT like what's an academic expert

and like what's their email?

Like who should I try to like make some more contacts?

But a lot of those are like early on in the process.

And I have a fairly hard line

where like I don't use any AI note takers out of respect

for the experts, the sources that I'm talking to.

Not only if I don't want their identity somehow leaked

if something happens with the software,

but also when I'm talking with an expert,

I don't really want them to feel like some AI is listening

to everything that they're saying.

I want 'em to feel comfortable enough to talk to me.

So I don't use any kind of AI note taker.

And also, my most analog thing I'm proud of

is I write literally every blog by hand.

Like I have my notepad

and I'm scribbling through everything.

So I think you can blend it,

but I think there's positives, there's negatives.

Try things out for yourself

and I feel like you'll quickly learn

what works and what doesn't.

I feel like the psychological research

supports what you're doing, Reece,

about like writing by hand, that there's something extra

that comes out of that process.

I have never written a story by hand,

and I so respect you for that,

and like you're like that much more amazing as a result.

We have a lot of questions that get at the core angst

of who's gonna be working, what does that job look like?

Urza9814 asks us this, I'm struggling to deal

with the shift in human skills being demanded

by the way my company is responding to AI.

Suddenly, managers think they can be coders,

which turned into every coder

being expected to be a manager.

I'm not interested in being a manager.

How do I find a job I'm actually capable of doing

in this environment when everyone expects

everything can be done by anyone,

exclamation mark, question mark.

Reece, maybe you have a sense of like, yeah, I don't know,

how should we thinking about this,

the manager, employee kind of divide.

Yeah, I do wanna make a note, though,

that like managers thinking

they can do your job better than you

is something that has always happened.

I've never experienced it at WIRED,

but as a young professional,

like managers thinking they can do the job better

than like the employee.

Like that's been going on for even more than just AI tools.

But I'm thinking about like how they're almost expecting

individual contributors to act

as kind of a manager of their agents

or like just have these like high-level skills of like,

I think one way that you can stand out

and really make yourself known

is leaning into the soft skills.

Like it feels like a cop-out but it's something

that is actually going to be helpful for you.

So whether it's having emotional intelligence,

whether it's being able to actually listen in a meeting,

and like respond like a human.

Whether it's like giving solid feedback.

Like these aren't the kind of technical skills

where I'm gonna tell you like you gotta learn to code,

because that's just like not the truth anymore.

You should have like a foundation

of whatever you wanna work at.

But this idea that they're now expecting you

to do even more, one way that you can succeed

is leaning into those human skills, I think.

Yeah, Kate, if you have any thoughts,

I'm curious to hear what you have to say.

What comes to mind for me

is just this idea that like we're in a moment

of crazy, fast transition and a lot of managers

have had some pretty shady and shaky ideas lately.

And this idea that everyone can be a manager,

and everyone can be a coder might just shake out

in a few months.

So like what I would suggest if I was your friend

would be just like, just hang tight for a second.

You know, like, yeah, this is probably really annoying

but it might not look like this in three months.

Like things are just evolving so fast,

and people are just like throwing spaghetti at the wall,

and trying new management structures.

Like I don't know if you guys saw Jack Dorsey

when he like laid off all those people at Block,

and he was like, you're gonna be reporting up

to an intelligence, it's gonna be a lair

that just like intelligence that you're gonna report up to.

And you're like, what does that even mean?

And like he's nuts, sure.

But you know, it kind of like speaks to this idea

that like no one really knows how to work with AI,

and it's kind of a messy situation,

and we're gonna be in a very different place

in like six months.

That's my personal take.

I don't know.

Kate, what do you think from observing the industry?

Okay, not to cheat,

but I'm gonna synthesize your two takes.

[Reece] Yes.

Very AI of you.

Yes.

What Reece was saying about soft skills is dead on.

Like in the age where everyone can allegedly be a coder,

being the personality hire is more valuable than ever.

And what you were saying is also spot on.

Like this is nuts.

It's not sustainable.

Half of the time, if not more than half of the time,

the push to make all of your employees AI-forward

automatons of their own.

Like it's not resulting in this huge productivity spike.

We haven't exactly like had a new Einstein emerge

from this moment.

So like everyone needs to calm down.

If you could hang tight, focus on your core competencies

and your personality, that is your best bet.

I do not envy anyone right now,

especially in certain markets

where there are these really concerted pushes

to make every employee an expert in coding, and Claude,

and blah blah blah.

Like it's rough out there.

And I'm also, you know, safely in the firm, cozy bosom

of Conde Nast at the moment.

So I'm not like, if I was job hunting

I'd be tearing out my hair and crying with people.

But yeah, don't like spend every waking moment

trying to learn how to do OpenClaw or something.

Unless you're a savant,

but like we're gonna move past this moment.

Yeah.

And kind of going back to the Block situation for a moment.

So if you didn't hear about what happened,

earlier this year, Jack Dorsey laid off a bunch of people

at Block, it's a financial company in the US.

And one of the big reasons that he was putting out

was because this was AI.

Like AI was so good.

They were having so much productivity and efficiencies

that they didn't need all these people.

And I think this goes back to questioning

what your boss says about AI,

and like what their motivations are,

because we were really early on that story.

I have a lot of sources at Block

who were feeding me a lot of different information

as that was happening.

And one of the big takeaways

from employees inside the company

is that like, sure, some of the AI tools are working fine,

like the AI tools were okay, but a lot of these job cuts,

according to the employees were actually due

to over-hiring during the pandemic.

So it wasn't exactly that AI was actually so efficient

that they no longer needed these people,

it was some other business decision

that AI was kind of running cover for.

So if your boss is telling you like this and that about AI,

just take a big grain of salt with that,

because, you know, in addition to things

maybe just being different in a few months, right,

as the markets change,

they probably aren't just telling you the full story there.

So I think approaching it with a high-level of skepticism

is gonna be a good option here.

So Block was one of the early ones

to have like a big and dramatic round of layoffs.

There have been several more since then, Oracle, Coinbase,

I mean the list goes on, Amazon had a bunch.

I'm curious if you kind of apply that same lens.

Like are those also only semi-sort of about AI?

Do you carry that assessment forward, basically?

I think in some situations, yes.

I think specifically looking at coders, software engineers,

it's one of the big industries that's being disrupted,

has been disrupted already over the past few years

with AI tools.

A lot of coders are using AI all the time, every day, right?

So I think it's specifically an industry

that there have been changes in.

And I think that has contributed to some of the layoffs.

But I do think there's kind of like, you know,

we talk about like greenwashing, where you like,

or Pinkwashing, there's kind of this AI washing

of like if your company isn't doing well,

and you need to do some kind of cuts, like it's AI,

AI is the reason.

Like I've not been reporting forever,

but I've been doing this for five years at WIRED now,

and there's been multiple rounds of layoffs

at different companies and the reasoning, you know,

it was pandemic layoffs and now it's AI layoffs.

So I think the, for software engineers, the job is changing,

but this idea that every AI layoff is actually due to AI,

I'm still skeptical.

So, you know, we're up at like

more than 140,000 tech layoffs this year so far,

which is a pretty big number.

Again, there are also a lot of openings

that people point to.

So it's not a clean story

of like the jobs are simply going away,

but anecdotally, there are a lot of people really struggling

to find a job.

And if we're getting this mixed messaging

of, oh, it is AI, it is not AI,

how should people like they be thinking about this moment,

how should they be re-skilling themselves, if at all,

you know, what is sort of like some advice for people

to navigate this moment,

given that we're not getting the whole picture?

I think that re-skilling

is definitely something we can't, you know, ignore.

I definitely think the wrong advice would be to say,

just pretend AI doesn't exist whatsoever.

Because if your job is telling you you need to re-skill,

you know, take it with a grain of salt,

but also, sign up for those courses.

Keep an open mind.

I know I was just talking about

how everyone needs to make their own decisions

about what ethical and moral uses of AI are.

Sometimes you might kind of have to compromise

if you disagree with these tools existence,

but your job really wants you to learn how to incorporate,

you know, Claude into your work somehow.

It's tough, but you know, we are at a moment

where it's important to play along as long as you can.

I wish I had better advice.

I really do.

'Cause I feel so much for people right now,

and it's sometimes so industry-specific.

It's like, I look at doctors

who have been sort of, you know,

there's all of this exciting stuff happening

with AI in medicine, but there's also this movement

to incorporate AI tools into like every step

of the diagnostic process,

including introducing AI note takers.

You know, like the world's best pediatric dermatologist

can kind of do whatever they want,

'cause they're so skilled that they don't have to worry

about their boss breathing down their neck

about not using an AI note taker.

I would give them very different advice

than I would give someone who's in middle management

at a big tech company like Amazon

where I know that they are absolutely required

to be experimenting with AI tools,

including Amazon's own tools.

Telling one, just do what you want, do your thing,

is good advice.

Telling the other that, is not.

So it's definitely a tough situation all around.

And if you're someone who is kind of caught

in the cross here too, where you're like,

I would never use AI at home, but my boss is asking me to,

and I need to use it at work in some way,

and you haven't done it before.

I think two little tips to kind of like make it less painful

is to do voice input.

So this might feel a little weird at first,

but especially, if you're in like an office setting

where it's not that odd to talk out loud,

speaking what you wanna say out loud to the AI tool,

hitting enter, and then having it output the text

is gonna just make that a little more seamless for you.

If you're using AI

and so then something you're not typing all day,

it's kind of how I've been using it for a while.

And then also, if you have to use AI,

and you're probably already well aware of this,

but watch out for those mistakes.

Like everyone talks about it over the past few years,

but it's gonna be really easy for you

to kind of get caught up in the conversation,

and follow along with what it's saying to you.

But just always have that critical thinking cap on,

super, super tight.

Yeah, the mistakes, I think,

are something we haven't really gotten into yet

and are so important.

Here's a question that kind of speaks to that.

User12345 asks the following, Most business information

is confidential, how can AI accurately silo

confidential information within an external to a business?

Okay, I think the answer to that is probably pretty huge

and it certainly is gonna be different

from industry to industry.

But yeah, how should we think about this?

Like question of confidentiality.

I know a lot of security teams out there

are losing their minds over AI and the risks that it poses,

and OpenCloud, oh my god, what a mess.

What have you been hearing and seeing?

It is such a mess.

It's such a mess.

Especially, you know, there was just a legal case

where a judge ruled

that, you know, when a lawyer inputs information

about a client into a chatbot, it voids the confidentiality

that lawyers usually have with their,

like attorney-client privilege.

That is a hot mess because as we all probably know,

lawyers are obsessed with AI.

Like that's why we're constantly seeing lawyers

get in trouble for filing cases with hallucinated citations.

So I mean, one thing I think if you are interfacing

with a lawyer at all, check and make sure they know that.

Like if I had to hire a lawyer for any reason,

I would ask them, how are you using AI?

Are you protecting the privileged information

that I am giving you?

And that sort of fans out,

and can, you know, apply to a lot of different industries

where, you know, data privacy is especially important.

Because some of this stuff is really unsettled.

Like there's murky legal questions

that remain up in the air.

You know, we've been reporting on all of these cases

where big tech companies have been sued

for their chatbots causing people harm.

And there's a lot of questions about where liability

for that harm falls legal,

like whether they can be held accountable

for what their chatbots have said.

A lot of companies would argue that the chatbots

should be protected by Section 230, which is this law

that provides shield for internet platforms,

that they're not responsible for the speech

that's made on their platforms.

So you get into questions of like, is chatbot output speech?

Yeah, huge mess.

I could talk about this for like four hours.

We have to do another one of these.

Yeah, that's a total mess.

And thinking about sensitive data, like lawyers, yes.

Like I think you should be approaching people

that you're working with, you know, like if I am the client

and they're the lawyer,

like ask them how they're using AI, yes.

But like also thinking back to like the medical situation,

don't be afraid to be a bit of a curmudgeon.

Tell your doctor, actually, no,

I don't want the AI note taker.

Like don't let these kind of,

if you're feeling a soft social pressure to go along

with, you know, an AI note taker in the doctor's office,

but you know, when you think about it a little more,

like you don't really feel comfortable with that,

'cause you don't know how the data is gonna be used,

it's a new startup, you don't know where this is going.

Don't be afraid to put your foot down in these situations

and have a little bit of confrontation

to say like, no, I'm here in person, take normal notes.

And if you're on the inside trying to use these tools,

and figure out how to configure AI for your company,

I mean, I feel like if you have any say over it,

try to limit what data AI has access to,

so it's not just ranging freely through your systems.

Like try and be a little more conservative

about what AI has access to using and processing.

And then also, you know, making sure

to redact sensitive information beforehand,

before you give it to an AI system

so that we can just try and like limit the exposure.

It just feels like we all have a little bit of a role

in making sure the AI is used well

and it plays out very differently

depending on who you are and your context.

Let's see, so you know, that was the business side.

What skills or activities

should non-tech business people pursue

to stay on top of what AI can do and the risks it poses?

That's kind of the follow on

to what we were just talking about.

Answer number one would be read WIRED,

but what else can they do?

Oh, man.

Yeah, read WIRED, obviously.

You know, I would be so generous to say,

read a variety of tech publications

to keep up to date on this

and play around with the existing tools.

Just don't give them confidential information.

Yes.

And kind of building on what we said earlier,

I think it's not necessarily about spending a bunch of time

learning how to do one specific software.

It's not about becoming a Claude master.

It's more about like staying flexible,

staying adaptable to new types of software,

new user interfaces.

And also, thinking about like non-tech business people.

Like what can they do specifically

is to get better at writing.

Like I think all this AI writing

has like actually increased the value

of like a strong personalized email, a message from someone

that I can tell they've thought about it,

and this isn't just some kind of slop hitting my inbox.

So I think it makes your communication skills

that much more valuable.

Totally, if anything, like just double down

on your humanity, what makes you unique,

what makes your mind unique?

Like these are the differentiators,

this is how you stand out.

One thing I really wanna talk about

that's kind of in between the lines

in a lot of these questions is like,

there's just been a lot of rage against AI, towards AI,

more talk about anti-AI protests,

more concern about like, if lots of people are out of work,

what the heck is gonna happen?

Like, I'm curious, because at the same time as you see that,

you also see people just adopting it,

because it's the ultimate cheat code, right?

It's like finally, there's someone

who wants to do all your work for you

and it just costs a few tokens here and there,

monthly subscription plan.

So are we gonna see people kind of become complacent,

and chill out, mellow out?

Are we gonna see more of a kind of backlash?

I'm curious what you see in the tea leaves.

More backlash, if I had to predict.

I ultimately don't think that the neo-Luddite movement

will successfully eradicate AI usage.

So I don't think that it's going to result

in like us abandoning these tools as a society.

I do think that we are not done with the backlash

and the rage, especially as like

lower skill entry-level work is the first to go.

We are watching college graduates

enter like a hostile job market.

A job market that makes like what I entered in as a grad

after the financial crisis look cushy,

and they're gonna be pissed, and yeah.

Yeah.

I see we have a live question from a WIRED reader

who says, EZ Recline wondered if AI built

on so much factual information

could ultimately restore facts and truth to society.

That's sort of a very different take.

What do you think?

Yeah.

Sure, Ezra, sure.

No, wait, so you're saying that like AI

could make everything more factual?

Like I'm not following this.

I'm thinking about like AI fact-checking maybe

on like the social platforms or, you know,

the answers aren't generally kinda often right

in broad strokes, so maybe sort of an escape

from the, you know, tiny subculture rabbit holes

that people can fall down where all of a sudden you believe

that like rocks will talk to you or whatever.

There was a really interesting piece that just came out.

The radio show Question Everything

worked with a news outlet that I forget,

everyone should Google it, that looked at research

that like shocked the researchers.

I don't know if you guys saw that,

it showed that people who believed in conspiracies,

who engaged with chatbots, that chatbots could successfully

like unblack pill people, basically.

They were like convincing people that reality existed,

and that things were true, and that was like not really

what the researchers were anticipating.

And the takeaway was heartening for people

who are concerned that the rise of AI

will lead to like a greater explosion of misinformation.

It showed that in some contexts maybe it could bring people

into a shared reality once again.

Obviously, that's one study and we'll need a lot more,

but I definitely read it and felt slightly better

about the world, so.

I feel better just hearing that right now.

I think some of my skepticism from this

is kind of the feeling of like the AI tools

are kind of what's poisoning the internet

with so much information

that you're not sure whether it's true or not.

So the idea of it could help bring some factualness sometime

is a little bit of comfort, right?

But it does kind of feel like

the oil company spilling a bunch of oil

then like cleaning up one little patch.

It's like, let's weigh both sides here.

Right, like, we got some facts, thank you.

We also lost a hundred thousand jobs, so, yeah.

So speaking on job loss though, like, last question here.

Liz wants to know, should we be pushing for more at work

to protect our jobs?

Should we be serious about unionizing?

Have we seen that work?

Where are we kind of on that?

I think there absolutely should be a push for workers

to take collective action and unionize.

Obviously, like anyone following the Hollywood strikes

for examples, has seen the limits to what Unions can do.

They can offer protections for workers,

but they can't eradicate the usage of AI in an industry.

So it's one step towards

making a more worker-friendly culture,

but it's certainly not a panacea or anything.

I think that pushing for universal healthcare

is absolutely critical right now,

because if we're seeing all this job loss,

it's like extra insane

that our healthcare is tied to our employment status.

So whatever we can do to push for healthcare to be universal

and uncoupled from our work, I think is gonna be critical,

not only for people being happy and healthy,

but also for people not having such intense backlash

that they actually do rush out with pitchforks.

So that's my big thing.

No, that's a really good point.

That's a good point.

I completely agree, Kate.

And to kind of just build on that just a little,

is that thinking about like your specific workplace

wherever you are, if you don't have a union,

unionizing could mean

that your union could set some boundaries

about how you're expected to use AI.

You know, what can be asked of you with these tools.

So especially if you're worried about

not only keeping your job, but if you keep your job

what it might look like in a few years,

a union could be an incredible advocate for you.

Thanks again for joining us today.

A recording of this live stream will be available shortly

right here on this page.

So you can come back and re-watch, catch moments you missed.

Reminder, please provide us with feedback and ideas

in the comments section below,

or you can email, [email protected].

Kate, you have a cool, new book project I wanna hear about.

Tell us a little bit about that.

Yeah, so I am so excited.

This is like a dream I've had for many years.

We're launching a WIRED book club newsletter in June.

Everyone should go sign up.

We already have the first book selected.

I don't know if I'm allowed to say what it is yet,

but it's really good.

Like I read probably 10 pages and I was like,

this has gotta be the first book club book.

We're gonna run a discussion

for I think like five to eight weeks,

and there'll be a Q and A with the author,

and it's gonna be a blast,

and there'll be extra book recommendations,

and you should all join.

Don't forget that as a subscriber,

you also get exclusive access to five premium newsletters

on topics ranging from the latest business news

out of Silicon Valley to the AI advances

shaping our world and the politics driving the country.

You can sign up at www.wired.com/newsletters.

Reece, Kate, thank you so much.

We've had a blast.

I have at least, can't speak for you.

I hope we get to do this again soon.

I had a blast.

I promise right now that I will never replace myself

with an AI avatar for these live streams.

Everyone is gonna get human, Kate.

Cool, I'll be checking in on you in 2040.

[all laughing]

I'm seconding that, and I know what the book is,

and I'm gonna go buy it right now after this live stream.

So I'm so excited for Kate's book club.

Support an independent bookstore.

Okay, bye.

Bye.

[upbeat music]