In-Ear Insights Generative AI and Job Losses

In-Ear Insights: Generative AI and Job Losses

In this episode of In-Ear Insights, the Trust Insights podcast, Katie and Chris discuss the impact of generative AI and job losses and its role in the corporate world. They debate the consequences of replacing human roles with AI, emphasizing the risks and short-term thinking involved in such decisions. The episode explores the evolving nature of content creation and the necessity of human creativity in the age of AI. Listen to gain insights on balancing AI integration with human expertise in the workplace, and the importance of staying adaptable in a rapidly changing job market.

Watch the video here:

In-Ear Insights: Generative AI and Job Losses

Can’t see anything? Watch it on YouTube here.

Listen to the audio here:

Download the MP3 audio here.


Machine-Generated Transcript

What follows is an AI-generated transcript. The transcript may contain errors and is not a substitute for listening to the episode.

Christopher Penn 0:00

In this week’s In-Ear Insights, generative AI has been taking over the world and has been taking over every conversation.

And we’re now starting to see hints of what I think is probably the worst case scenario for generative AI where companies are making very large decisions about general AI, to the detriment of people to the detriment employees, I was doing some work over the weekend, writing some code to process news.

And because we’d like to take a newsfeed to find content worth sharing.

And I noticed a very big trend in the last seven days, which is a lot of companies, particularly big tech companies saying, Hey, we’re doing all these layoffs because of generative AI.

Google, the most prominent of them.

One headline, which is very interesting was that Google has terminated his contract with the company app.

And that has provided the search quality ratings guidelines testers, the human testers for the last 17 years, these are the people who help Google’s to help create training data for Google’s algorithms to say, this is a good quality website.

This is a bad quality website.

And the speculation have to be clear, this is a speculation, there’s been no official word is that Google is going to use general AI to do those tasks instead to evaluate web pages, and to create them.

But we have a lot of news about all the different layoffs and how company executives, particularly senior executives, are looking at general AI, as a replacement for employees, another big tech company said they’ve removed about 80% of their marketing department, but because they don’t need it anymore.

They have good solutions that create better quality work than they were getting out of their humans.

So Katie, given this a slightly alarming trend, what’s your perspective on it? And how are you thinking about the use of generative AI as a senior executive?

Katie Robbert 1:52

You know, it’s it’s so dependent on the situation at the company.

But as a blanket statement, I think it’s a mistake to wholly replace humans with generative AI.

And I say that, not in that generative AI can’t do what humans do.

We’ve already we’ve discussed it, we’ve proven that it can.

But I feel like it’s short term thinking, because the technology is still so immature.

The technology is still changing on a daily, even like hourly basis, that it’s not stable enough to feel calm, that anyone who feels confident that they haven’t mastered and that they can replace humans with it, I feel is fooling themselves.

Let’s just take the example of Google.

Google obviously has a lot of resources, way more resources than you and I will ever have.

And they’ve probably probably been working on this for a long time.

I still feel like it’s a mistake to replace 100% of your human raters with generative AI because of how immature the technology is, when I did clinical research, we actually one of our studies was to build a rating algorithm.

And it started because you have to start with the human as the control.

And then the tech was the experimental.

And we did all of the same tasks that the you know, computers did.

And so basically, what we were doing was we were raiding substance abuse forums for you know, potential of people talking about use and abuse of opiates and stimulants.

And we almost never had 100% agreement between the humans and the machines.

Now, the machines, understandably, had more consistent results.

But they were limited, and they didn’t take in new information very well.

Whereas the humans were able to adapt faster.

Now, this is, again, this is going back maybe 1015 years, this was much older technology, but it’s just to me, it’s reminiscent of, you’re never going to have 100% human and machine agreement.

And so to say that machines are just going to do it now, to replace the humans to give the humans the information they’re looking for, to me, is a huge red flag.

Christopher Penn 4:20

I can definitely see for more complex jobs.

That’s absolutely true.

You know, if you look at the work that we do on a day to day basis, the the biggest value add is the human, right, because we have machines that do all the computation, all the data crunching, even all the basic analysis, but you still need that human to say okay, he will, here’s what this means.

And here’s how it fits into the big picture context of your organization.

Right? There’s when I think of some of our largest clients, they have very complex organizations, they have very complex interpersonal dynamics.

And so whatever the machine spits out clearly needs some additional context to make the recommendations pal audible.

But when you take a task like search quality weighting guidelines to me, that is a, that is a job that is a task of that one tasks.

Yeah, you can automate that.

Because it is literally that’s all you do it the day in and day out, you just follow this little checklist and you go down the checklist and you fill in numbers, that clearly is a a job.

That is the task when we worked at the old agency, there was one job on staff, besides getting coffee for people, that job was to copy and paste links from, from Google search results into a spreadsheet, eight hours a day, I don’t know how this person got these fields in claw their own eyeballs out.

But that is another case where that job is just a task.

I mean, you know, people get their own coffee,

Katie Robbert 5:41

it’s gonna say no one ever got coffee for me, I was in the wrong role, apparently.

Christopher Penn 5:45

And those, those jobs are, I think, at substantial risk.

And so the question is, how many of them are there? And what do we do with all those extra people, I’m

Katie Robbert 5:57

sure if I asked you, you would say some sort of apocalyptic plague to, you know, thin out the herd or something like that.

But in realistic terms, you know, I hear what you’re saying, you know, the search rating guidelines, it is a checklist it is, and I don’t disagree with you, that is something that you can absolutely give to generative AI, where I see you still need human intervention is that people evolve, I was actually talking with a friend about this the other day about the evolution of content, and specifically one particular content creator, and the content creator, is continually reinventing their content to make it fresh for new audiences.

And so we were sort of talking about it in the sense of, Well, it started, you know, here in this, you know, form and now, here are all the phases that it’s gone through, and it still feels fresh today.

And my concern with giving generative AI this task is that it’s not then also picking up on the nuance of the evolution of how we, as content creators, are evolving our content to different styles, different platforms, different, you know, delivery vehicles, to reach new audiences.

And so I just, you know, I’m just sort of saying that like, as a blanket statement, obviously, it shouldn’t forever, like, didn’t do it, the algorithm to keep up Blah, blah, blah.

But I feel like the point that I’m trying to make is that, as humans, we’re constantly trying to reinvent the work that we’re doing to make it new and fresh and better.

And wouldn’t you also want human intervention to pick up on that, because the thing that we know that AI is not good at is nuance

Christopher Penn 7:39

Yeah, out of the box is not good at nuance, because nuance requires understanding of very complex situations, which requires a lot of data, you can give the machines, some level of it, if you add that relevant data to their latent space, so that they, they can understand it, and provide very clear instructions about how to handle it.

But that requires advanced skills that a lot of folks were using general AI, don’t necessarily have shameless plug, if you’d like those skills, go to AI slash AI course, where you can take a course to learn a lot of these basic and then later on advanced skills.

I think, though, one of the challenges and I don’t have any information right now on this is to try to figure out just how large that body of humans is.

For example, one of the things that has been the case up till now is a system called reinforcement learning with human feedback where companies like Google and open AI and meta hire contractors, 1000s of contractors to basically take spreadsheets and have humans annotate answers, like here’s a question, Here’s the correct answer.

And that creates trained datasets.

In the last six months, we have seen an explosion of reinforcement learning with AI feedback, where machines are essentially are scoring and grading each other.

And, to some surprise, in the industry, reinforcement learning with AI feedback is better than humans, it does a consistently better job to the point where you now have larger synthetic datasets and you do human datasets.

And those synthetic datasets are higher quality.

So you are getting essentially better models from that.

Given that backdrop, how many? You know, again, it comes back to what do we do with the people? What do we do with? If you think about every company, there are plenty of jobs, which are maybe a couple of tasks.

And I don’t think what’s happening with a lot of these companies is they’re just wholesale filing, firing your person saying like, hey, you know, your job is completely redundant.

We’ve got a machine that, you know, here’s the machine you’re out.

What I think it is, is extraction of roles where you say like we had 80 writers on staff, but now we only need 1010 People work ChatGPT And so we can get rid of 70 we have not gotten rid of the humans, the humans are providing oversight, but we for sure don’t need all 80 We can get away with 10

Katie Robbert 9:59

So Before I get into that, I want to make a comment on what you were just talking about in terms of the, you know, human annotation to the questions.

I am unsurprised that generative AI is doing a better job because humans ask the same questions over and over and over again.

And there’s only so many answers to those questions, you know, the technology might change, the answer might get updated.

But it’s all the same questions.

And the thing that generative AI has over a human is basically an encyclopedia at their fingertips.

Whereas we, as humans have to like, figure out do we know the answer to that question? Can I go look it up? Can I trust you know the source.

So I totally, I’m unsurprised that generative AI has been able to do a better job of that it just again, it has more resources, and can do it faster.

But humans don’t ask different questions.

We ask the same questions over and over and over again, just to maybe like phrased differently, if you’ve ever looked, this is a side note, if you ever looked at like a local Facebook community of like, one of the town’s people ask, is the trash getting picked up? Is the trash delayed? When is the trash getting picked up? Was my trash picked up? Was your trash picked up? And without fail, the admin will pop in with a very passive aggressive, you know, photo of the search bar in the community group and say this is the search bar, go this question has been asked 100 times already go look for the answer before you ask it.

And so that’s sort of what it reminds me of is that these machines take away that annoyance of humans asking the same questions over and over again.

Now to your question about what do we do with the humans? I feel like, you know, you’ve talked about how we need to now this is the opportunity to tap into our creativity.

So in that example, have you had 80 writers, generative AI does a pretty good job of writing for you now.

So you can let go of 72 Left hand those 70 writers should be less focused on, you know, the technical aspect of writing and more focused on the creative ideation of what’s not there.

So if I had a team of writers and I knew that generative AI could write the majority of what they’re doing, I would challenge them to say, All right, now your job is to take a look at what’s going on outside of the walls of this company.

What are people talking about? What do they want to know about? What are the gaps? What are people not hearing enough about? And can you start to put together those content plans? Or what are the types of content? What are the delivery methods of content that we’re not doing? And what does that look like for us to do that? So it’s really an opportunity to say, okay, get out of the box of just writing the thing that we tell you to write, and go find out what’s new, what’s next? What’s not being done? And bring that back to the table?

Christopher Penn 12:59

Do those people in this example do those people have those skills? They’re writers today, they make they make blog posts and stuff.

And now we’re almost asking them to become in many ways, researchers?

Katie Robbert 13:12

That’s a good question.

And I don’t know the answer to that question.

And I would imagine that maybe, you know, a percentage of them do.

And the ones who don’t either need to figure it out, or find, you know, they’ll probably have to find different employment, because what is now being asked of them is different.

And that’s going to be a challenge for a lot of employers.

And so in this example, you know, I’m a really bad employer, who I’ve hired you to do one job, and now I’m asking you to do something different, that may not be within your skill set, or you’re probably going to get let go.

And that’s really, you know, shame on me for not thinking longer term, about where I want the generative AI to be, and where I need the team to be.

Because if I were thinking long term, I would have started these conversations with the team a while back and said, You know, I feel like what’s going to happen is generative AI is probably going to be able to take over these tasks.

And so what I need you the human team to be thinking about is the following three additional skills, you know, and here are the resources that I can provide to you to skill up to get to where I’m going to need you to be, you know, in the next 612 18 months, you know, and so that’s the way it should work, but I 100% guarantee that’s not the way it’s working.

Christopher Penn 14:31

I think though, you just said something super, super important.

I think it’s probably the most important part of today’s episode.

Whoever you are, as an employee, you need to be taking an inventory of what you do, and saying, how much of this can AI do? And if you don’t know the answer to that, it is time to sign up for the AI service of your choice.

And try to try your best without, you know, where you exist.

Doing ideas to make the machine do your job or as much of your job as possible, because that will give you a sense of risk assessment to say like, here’s, yeah, a lot of my job can be done by machines.

So what value do I need to provide to keep that from happening? And there are a lot of people I see this particularly on threads, there are a lot of people who are have their heads very firmly in the sand and saying, you know, what, no AI for any of this stuff.

Like, when I look at the news headlines, as it was, which inspired this episode, a lot of companies going, Hey, we can save a whole bunch of money because people are our biggest expense.

Even Trust Insights I was looking at our financials from last year over the weekend, people are our biggest expense, if we can reduce the number of people speaking at large corporations, if I can reduce the number of people, I can cut my expenses significantly, and still deliver a lot of value.

In our case, we look for as a small business we look or how can I get gender AI to do even more so that I can increase value without adding a ton of headcount? The critical thing for everyone to be doing is how much of my job can be done by machine.

And then what is my plan as a person with a career to to augment AI, because companies will make in this case, you know, profit based decisions.

And that net profit based decision is reduced expenses as much as you can.

Katie Robbert 16:26

When I think about past experiences, and past jobs, the biggest expense wasn’t individual employees, the biggest expense was the employees were the teams that sat in meetings all day every day, and then had to try to get things done after hours.

And so it’s the meetings, that if you add up the people in the meeting, you add up their hourly rate, and then you try to figure out what they’ve accomplished.

That’s your big, that’s likely one of your biggest expenses.

And so if you are someone who just sits in meetings all day, either taking notes or pontificating, or, you know, just listening to someone, you know, a big ol windbag, and you’re like, why am I here I have so much work to do.

That puts you at risk.

Because, you know, meetings are something that are super easy to get rid of, or to cut down on.

You know, Chris, you’ve you’ve sort of shared this advice before, like, if you can look at your calendar, and see the same meetings or tasks or repetitive, and you know exactly what’s going to happen in your work day, for the next seven days, you’re at risk, because your job is so repetitive, that it’s really right for something like generative AI to take a lot of those tasks.

Thankfully, my day, I never know what the heck is going to happen.

So I’m not at risk, yet.

I’m working on putting myself at risk, that would be a lovely thing to have happens to have generative AI take over a lot of the nonsensical things that I feel happened during the day.

I’m just not at a place where I feel like it’s repeatable and repetitive and consistent.

That’s my goal, though, I would love to have generative AI take over a lot of what I do.

But we’re a small company.

So we do a lot of things, we wear a lot of hats.

And so for me, I would love to see it happen so that I can focus on the things that are more important.

And those things that are more important are what’s the future of the company, let me build relationships so that we network, let me do the human thing.

And I feel like I’m not at a place where I’m doing enough of the human thing.

I’m still doing the things that the machines could be doing.


Christopher Penn 18:35

as you mentioned, these tools are evolving so quickly, and, and gaining capabilities on a regular frequent basis that are beyond what we could have imagined 18 months ago, that day may be sooner than then than we think.

And certainly it’s going to be one of the things where you won’t notice until it’s it’s done, right is you chip away at it until like, hey, you know what, I actually am getting to do more of the work that I want to do because the machines are handling a good chunk of it.

For us, that’s great.

That’s what that’s where we want to be to be able to spend more time doing stuff that’s valuable for companies that are motivated more by short term profits.

I think that poses a much bigger existential problem for a lot of people who work at them.

What, what do you say to those folks who are like, Yeah, I’m, you know, I was a mid level manager and I’ve been been counseled out because that role has been eliminated.


Katie Robbert 19:33

starting to get into the psychology of humans.

No, and I say this because it’s a big topic and, you know, to let me see if I can sort of make it, you know, concise and summarize it.

There is a lot of fear.

With achievement.

There’s a lot of fear that comes with trying to succeed because with that comes failure.

So I know a lot of people in my you know, Personal circles that are very, I wouldn’t say happy, but are content to be middle managers, because there’s very low risk involved with that kind of a role.

You don’t have too much authority, but you have just enough authority, you don’t have a lot of responsibility.

But you can do stuff if you’re asked.

And so it’s a very complacent role to be in.

And they choose those roles, because they’re risk averse.

Those are the people who were at biggest risk for if and when generative AI or some other tech takes that role, they’re not going to know what to do, because they’re so risk averse, because, you know, focusing on what they want to do, what their passions are what you know, fulfills, that makes them happy.

It’s just never been a priority.

And so that is, the advice that I would give to those particular people is, you have to get comfortable with being uncomfortable.

You have to start thinking about, what is my life look at if this job doesn’t exist? And it’s a really big question, it’s a hard question.

It’s something that, you know, to be completely honest, some people may have to work with, you know, a professional counselor to figure out because there’s some sort of a mental block, because taking risks with their career is going to be anxiety inducing, it’s not something that everybody financially can afford to do.

And so if you’re not thinking about, what is my career look like beyond this particular role that I’m in, you have to start somewhere, even if it’s, what are my hobbies? What do I like to do outside of work, so stop thinking about it in terms of its work, work work, only, maybe you’ll find, you know, your second career, outside of what you thought you were meant to do outside of paycheck to paycheck, maybe there’s other things out there, but you have to start thinking beyond what’s happening today.

Christopher Penn 22:05

I think you just nailed it, in terms of how people need to be thinking about evolution is the things that you are good at professionally, and the things you are good at personally, may not be the same thing.

And they probably aren’t the same thing.

But increasingly, you need to to be looking at Can I do that thing that I do personally, because it’s probably going to be of more interest to a subset of the subset of the audience than what you do professionally.

People are not going to be super thrilled, probably.

There’s, there’s something for everyone.

But spine, I’d be super interested in learning more about project management from you.

But if you have have an interesting take on coffee preparation, or painting or something, increasingly, we see people making literal careers out of that I was doing some work over the weekend, just looking at some different YouTube channels and things and there are a lot of people who are making a lot a decent living on stuff that like hmm, that would not be that would not have been a career that our guidance counselor told us about 30 years ago, like, Okay, you’re going to be a YouTuber who does professional reviews of x.

And it’s like, that’s weird.

And, you know, making 50 $100,000 a month on sponsorships for your channel.

So there is opportunity, but it is a big change.

And for a lot of people, like you said, who are not mentally in that place yet.

This might be the wake up call that yeah, you need to start thinking about that along those lines.

You know, what is your side hustle is as much as I don’t love hustle culture, and things.

Having that.

I guess life raft, if you will, the emergency backstop.

It’s not the worst idea.


Katie Robbert 23:50

You know, it’s people may not mentally be there, and people may not financially be able to be there.

But you should at least I would encourage everyone that you you have the time and the mental capacity to make a list.

Make a list of actually, this is an exercise.

And I always think back to this, that one of my old bosses was doing when she sort of felt stuck in her career.

And she went from being the director of the PMO to a postpartum doula.

And you’re like, how, how does that? How do you get from one to the other, but when she walked through her logic, basically what she did was she got some of those big old, you know, pieces of sticky board paper that she just put up in her walls in our kitchen.

And she just started to make lists.

She made lists of things that she likes to do.

She made things of listed she doesn’t like to do, and she didn’t think about it in terms of career.

She just thought about it in terms of herself as a person.

And that’s how I want people to start thinking about this exercise is forget you know, you said these aren’t the you know, jobs that are guidance counselors.

Forget that forget about what the job itself the construct of the job looks like.

And think about yourself as a person and what you enjoy doing.

And what you don’t want to do.

There is a lot of I see a lot of ads for, you know, remote work you can do that doesn’t require you to talk to people.


Put on your list.

I don’t like to talk to people because I’m introverted or No, I mean, it, I overdid.

And so if I chose to start my career over, I would probably choose a career where I was less out in front, and I was more I could just do my thing, and then live my life.

But why do they have to be separate things.

And so if I’m making this list, I’m gonna put on my list.

I like to be outdoors.

I like animals, I like to, you know, think about plants.

I like to be creative, I like to problem solve and organize.

And so I can start to see those pieces come together.

And I can start to see those patterns.

And the things that I don’t like to do would be on the other side, and I would just keep building out these lists.

And, you know, until I felt like, you know, what, there’s something here, I feel like, maybe I should do is x and maybe start pursuing that maybe I need more education.

Maybe I need to find new networks, new communities, communities are a great place to start right now, if you’re not even sure.

You know, can I start over in my career, what’s happening? Join a couple of communities a really great one is our free slot community analytics for marketers, you can find that at trust for marketers, and people are just talking about their lives beyond just analytics and marketing just every single day, I feel like I learned something from our community all the time about what people are doing.

And I’m like, huh, I never knew that was a thing that I could do, or that’s really interesting.

I want to learn more about that.

And that’s a really great way.

So that you’re not just taking this all on by yourself and you’re isolated.

But join a community virtual community, you don’t have to see people, you can just you know, stay in your home behind your keyboard.

But it’s a great way to get new ideas about things that you didn’t even know were possible.

Christopher Penn 27:14

And I would add to that, as much as we started the show, talking about the somewhat short sighted decisions that people were making the generative AI.

This is also a good application for generative AI, do those lists, do that exercise, make a comprehensive list? Here’s what I love, because I don’t love and then ask all the machines.

Hey, here’s what I love and what I don’t love.

Let’s talk about career options that maximize what I do love and minimize what I don’t love.

And you know, no restrictions on industry, no restriction on Well, what are my options? Because, again, these tools are basically walking encyclopedias.

Right, have a comprehensive knowledge of very large space.

It’s a great private way to ask those questions.

Maybe you’re not even a point where you are, you’re in a situation where you have trouble trusting other people ask a machine and a private context, say hey, here’s what I’m thinking of it, you might be surprised at some of the answers that come out.

So in the short term, do your career and Ettore in your in the short term, do your job inventory to see how much AI can and can’t do of your job? Be realistic and honest with yourself and to the extent that you can look for ways to either incorporate AI into what you do so that you are ahead of the trend or be building that life raft so that you have an option.

As Katie mentioned, if there’s a challenge, if there’s a place that you’d want to come talk about it you feel comfortable doing so hop on over to our free slack group go to AI slash analytics for marketers where you have over 3000 other marketers are asking answering each other’s questions every single day I saw a bunch popping up on the screen on my second screen while we were chatting and wherever it is you watch or listen to the show if there’s a challenge rather have it on instead, go to trust podcast where you can find us on every major podcast platform and while you’re on the platform of your choice, please leave us a rating and a review.

It does help to share the show.

Thanks for tuning in.

I will talk to you next time.

Need help with your marketing AI and analytics?

You might also enjoy:

Get unique data, analysis, and perspectives on analytics, insights, machine learning, marketing, and AI in the weekly Trust Insights newsletter, INBOX INSIGHTS. Subscribe now for free; new issues every Wednesday!

Click here to subscribe now »

Want to learn more about data, analytics, and insights? Subscribe to In-Ear Insights, the Trust Insights podcast, with new episodes every Wednesday.

2 thoughts on “In-Ear Insights: Generative AI and Job Losses

Leave a Reply

Your email address will not be published. Required fields are marked *

Pin It on Pinterest

Share This