In-Ear Insights: AI Will Take Your Job

In this week’s In-Ear Insights, Christopher Penn and Katie Robbert discuss the impact of artificial intelligence (AI) on jobs. They explore whether generative AI will take over jobs, especially for junior members of an organization. They discuss the aspects of jobs that AI can and cannot replace, such as repetitive tasks versus human creativity and accountability. They also examine the potential management and leadership problems arising from the inverted pyramid of employees that AI can replace, and the future of junior-level employees in organizations. Ultimately, they suggest that individuals should identify their interests and skills and explore how they can supplement and collaborate with AI, rather than seeing it as a threat to their job security. Watch the video to learn more about the evolving relationship between humans and AI in the workplace.

Disclosure: summary generated by AI from the show’s transcript.

[podcastsponsor]

Watch the video here:

In-Ear Insights: AI Will Take Your Job

Watch this video on YouTube

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, let’s talk about artificial intelligence, particularly generative artificial intelligence, and whether or not it will take your job and in the headlines and you look at how it’s being deployed, the answer increasingly seems to be, I will take your job, or at least a good chunk of it, particularly if you are more junior in an organization, although the technology is applicable to every level of the organization.

So Katie, how’s it looking for the CEO Trust Insights? Is AI going to take a job? Am I going to be reporting to Katy GPT anytime soon?

Katie Robbert 0:38

You know, it’s interesting, I think, in some ways you will be, I think there’s certain aspects of the job that I do, that artificial intelligence will take over.

So we were having a conversation with a colleague of ours last week, and he was asking if generative AI could create a crisis comms plan.

And so typically, those kinds of plans are created by more senior team members.

And the answer is, yeah, with enough information, then the artificial intelligence can create a plan, which means it can also create a marketing strategy, a business plan, all of those things that are typically reserved for, you know, the big brains of the company, I’m putting this in big, big, big air quotes.

Because it’s really, it’s not big brain stuff.

But the thing is that we typically think that only executives can do, artificial intelligence will be able to do so then what comes up is, well, what’s left for the humans.

And it’s, it really depends.

And so if we take that example of a marketing strategy, the things that we would need to feed artificial intelligence are, here’s the data, here’s the goals that I want to hit, here’s my KPIs.

And here’s about you know, my industry, my business, you give all of that to generative AI, and it’s going to give you a marketing strategy.

And so I can just go, you know, sit in my cabin, and let AI take over.

But what’s missing? Is that nuance what’s missing? Is that true insight into really understanding what my business my team, my clients are capable of and expect, making sure that I’m the one who’s, you know, keeping up on what’s changing, because human behavior changes, you know, with the snap of your fingers.

And you have to give that information to the machine in order for that to be you know, updated.

So, I think, yes, there’s gonna be aspects where Chris, you will be, you know, getting your information from artificial intelligence, but then, you know, you can’t get rid of me that easily.

You can try.

Christopher Penn 2:57

It’s interesting, because over the weekend, I was helping out a friend.

It live lives in works in Thailand, and needed a independent contractor contract written up now I know US law pretty well, I don’t know one thing about Thailand law about Thai law, and stuff.

But the ChatGPT model came up with the things that are and are not enforced in that country’s law and stuff.

It’s like, oh, okay, well, here, here’s a good first draft, and then you obviously would get handed off to a real lawyer.

And I think that’s, I think that’s where the human beings, the human beings still play three important roles, right.

One is relationships, there is a lot to be said for talking to another human being certainly everyone found out during the pandemic, how important that is.

To there is, as you said, the uniqueness right the the ability for people to create something as as net new, that has not been seen before.

Now, granted, a lot of our work is not right.

If it if you’ve done it more than three times a machine can do it, right? If it’s a template, if it’s a contract this time, the other thing that’s yeah, that’s repetitive work that a machine can do that.

But coming up with the exact wording that you want to use in a in a chapter of a book.

Yeah, you’re, the machines are not gonna be able to pull that off.

And the third thing that I think we we tend to forget about is accountability.

The human being is accountable, right, you can be taken to court, you can be fired, you can be demoted, you can be promoted, you can be given a bonus.

That is not true of the machines, the machines are owned by, you know, presumably a corporate entity of some kind.

But at the end of the day, even with stuff like copyright, the US copyright offices said works created by machine are ineligible for copyright because copyright only goes with the human being.

So if your job I would say if your job contains one or more of those three major aspects, it’s gonna be hard for a machine to do that.

Here’s the downside.

The vast majority of jobs, do not have strengths in more than one of those areas if they have the strengths at all.

I think back to our Time at the PR agency, watching an account coordinator is copying and pasting eight hours a day from one spreadsheet to another like yeah, that’s, that’s that’s not a, there’s no uniqueness.

There’s no accountability needed.

Right.

And there’s no relationship.

Katie Robbert 5:15

Right? Well, and so, you know, it’s interesting because this question will AI take my job has been around, you know, as long as the technologies existed, you know, we’ve been asked this question, hundreds of times we’ve spoken about it, and it really comes down to AI won’t necessarily take your job, AI will take aspects of your job.

And in some ways, you should be grateful for that, because it’s generally going to be the repetitive, mundane things.

And what it should be doing is unlocking that time for you to do the deeper insights to do the relationship building to really think through the future planning and what’s next.

And if you’re not looking at it that way, then you’re definitely probably anxious that you know, AI is going to take your job.

But you don’t have sort of that backup plan.

So to your question about, you know, should a CEO be worried? I’m thrilled.

I think there’s aspects of my job, that it is repetitive.

It is, you know, it could be more automated, but it would give me time, to have a better viewpoint of everything that we’re doing and think differently about how we’re productizing how we’re positioning the services that we offer.

So for example, you know, Chris, you’ve been doing a lot of work with prompt engineering, everybody is coming out with, you know, a PDF, or an ebook or the top 50 most effective prompts and blah, blah, blah.

So, okay, what does that mean, for us? We also do that work.

So how do we position ourselves? How do we differentiate? How do we add different value with prompt engineering, and that’s the kind of thing that I don’t see generative AI being able to handle really well.

But it is still very much needed, because what’s going to happen is, you’re going to have a lot of generic and similar information going out there because it’s machine generated, Oh, absolutely.

Christopher Penn 7:17

And a lot of the content that even people are putting out is put out without necessarily having a lot of knowledge about how the systems work.

If you’d like to get ours you can get it’s free.

It’s a one page PDF, zero cost, no strings attached, no form to fill out, just go to trust insights.ai/prompt sheet, and you will get our guide, at least for the ChatGPT API.

The thing that I see as as a major concern is is actually a management and leadership problem.

If you have, you know, we’ve we’ve worked with worked with and in agencies and very hierarchical organizations and the people at the bottom of the pyramid, very much like 80 to 90% of the job a machine can do, because that’s like the accountant coordinator, you’re copying and pasting today, right? Or you’re you when you have to write a press release for a client, what do you do, you take the last press release, you did you copy, paste it into a new document, and you change the names, and then you may be changed the CEO quote, but functionally, it’s something that machine can do.

So the pyramid almost becomes inverted in terms of the number of employees that you need in those layers, right? If you have 100, people writing rough drafts of content today, you probably need like five of them write, if you have, then you have a 50 people editing the content, you probably need 10 of them.

And what this sets up is two things.

One, it creates a very large employment problem for those junior levels.

And B, it creates a leadership problem because your future leaders come from those ranks, right? If there’s five people left out of 100, your pool of eligible candidates has shrunk by 95%, which means that in five years time or 10 years time, who you can choose from for directors, VPS Cornett office folks, that pulls real small and you may not necessarily they may not be the right people, because we’ve certainly seen plenty of cases where, you know, someone’s technically very competent, but poorly.

Katie Robbert 9:27

So it’s interesting.

So if I were in that position again, knowing that AI will take my job, I would start thinking about what are the things about my job that interests me and you know, I like the people management but I also really enjoy, you know, pulling things apart and understanding it, you know, completely and so I would say Alright, so I’m in this, you know, junior level position.

I don’t have a lot of authority, but I need to make sure I protect myself down the line.

So what can I do about that? Well, you If I’m not a strong people manager, then maybe my skill set is better suited at, you know, learning how to run the machines, learning how to operate them, building the processes, in which, you know, companies will use the machine and human.

So what does that look like? Together? Because you need both.

And so there’s definitely ways to think about, you know, AI taking your job, but how can I then start to future proof my job leaning into the skills that you have that you’re really interested in, you know, maybe research is your jam, but AI is going to do some of that.

You can supplement the research that AI does, by fact checking by, you know, making sure that the data coming out isn’t biased, or those kinds of things like there’s ways to think about it, because those are jobs that humans will still need to be doing.

If you know, pupil management is your thing, well, great, there’s going to be plenty of need for that, as you’re going to have this whole new class of prompt engineers and people who are running AI for companies coming in, someone still needs to wrangle all of that and make sure that things are prioritized, make sure communication is happening, you know, all of the things that a manager typically does.

So there’s ways to think about it.

But I would first start with what interests you about your job? And is there a way that that works into all of this new technology.

The other

Christopher Penn 11:29

thing that’s going to be really challenging and is an urgent problem for entire class of businesses to solve is the way AI works.

Here’s the list of businesses, law firms, accountancy fees, PR firms, marketing firms, ad firms, what do these all have in common, they all build by the hour, right, we build account coordinators time at $100 an hour, and so the clients retainer is $20,000 a month.

And to get X number of hours.

If your account coordinator now takes five minutes to do a press release, instead of the six hours that it takes them to struggle, yes, you’re gonna get see massive efficiencies.

But you can then no longer ethically build that client for those six hours, right? If it took five minutes, it took five minutes.

And so if your law firm is used to billing in six minute increments to do an employment contract, and GPT-4, spits out, basically a 95%.

Done, contract, the $3,000, you would have charged for you know, five hours of work is now five minutes of work.

And you can only bill, you know, $30 or $80 for that thing.

And I don’t think the management of these hourly based firms have figured out this is a business problem needed to solve one of the smartest things you ever had us do, as a company is the day Trust Insights was founded, we are not billing by the hour for anything we are building because you are paying for the 30 to 5070 years of collective knowledge that we have, that’s what you’re paying for.

And I don’t think a lot of companies aren’t in a place to make that pivot.

Katie Robbert 13:09

Well, and it’s, it is going to be a really steep learning curve on how to do that.

And so, you know, let’s go back to that example of, you know, putting together you know, employment contracts, or whatever the example was, yes, you are creating it in a fraction of the time.

However, if you have a good, you know, business sense, you can scale that and create, you know, you know, versions of it, and you know, get it out to like a much higher quantity, versus right now, maybe you can only take on two clients, you can now take on 200 clients.

And so figuring out what that looks like to make up for, you know, the lack of time that it now takes to put these things together.

I was talking with a friend of mine over the weekend, and she works as a paralegal for an attorney.

And just the things that she’s describing in terms of the preparation and the amount of time things take and just like you set dates, just to set dates within the court, like artificial intelligence is going to handle a lot of that, but some industries just are not there yet.

And so it’s a good opportunity for someone who’s thinking about, you know, future proofing their career.

Wouldn’t it be great if you were the one who started to bring those processes into, you know, your agency, your law firm, whatever it is knowing you know, there’s going to be negative side effects in terms of like billable hours, that kind of stuff, but then you become the expert for not only your firm, but then like, Oh, hey, the firm down the street needs the same thing.

The firm you know, three doors down also needs the same thing and you become that go to person to help introduce this into their processes.

Christopher Penn 15:00

The Law Firm example is a really important one, because it’s also going to create a massive disparity even just inside the courtroom, think about if you’re a large law firm, and you’ve got the resources to, as we talked about in last week’s live stream, the ability to build a custom fine tuned model.

And maybe you incorporate all of case law that’s publicly available into this model and tune it.

Imagine going to trial on one bench is a lawyer and their client on the other bench is a lawyer and a paralegal with a laptop with the your custom large language model there.

And every argument that is brought up, you are able to refute immediately with the entirety of case law at your disposal, and the ability to construct those arguments very, very quickly.

Because you are referencing stuff faster and better, in more complex ways than the other attorney ever could, right.

So there is an opportunity for your firm to start beating the pants off of other firms, because one thing that we’ve been saying for quite some time now is the professional with AI is going to overwhelm and take the job of the professional without AI AI is not going to take your job, the professional with AI is going to take your job, if you are not that professional.

Katie Robbert 16:15

It makes me think of other jobs that have just been so highly specialized.

So think about, you know, a court stenographer, they have those, you know, special machines that are like three or four keys and you start to learn the shortcode of, you know, how to basically transcribe live, what people are saying, well, I mean, we use transcription tools all the time.

And so is that a job that people go to school specifically for and get highly trained on, that’s not going to exist in a few years, probably, you know, you have the medical transcriptionists, who take all of the recordings that you know, a doctor sitting in the corner with this little microphone talking through his cases, and then someone manually transcribes all of those, sure, someone’s still gonna need to edit those.

But the bulk of the transcription can be done by a machine, as long as that industry adopts it and knows that those are going to be jobs that are going to be changed.

So I think in that sense, AI will take those jobs very easily, which is unfortunate, because it’s going to put a lot of people who only have that set of skills out of work.

Christopher Penn 17:25

And to your point earlier, there is an opportunity for a small minority of folks within that field to become the ones that train and tune machines, right? If you are a if you are a medical transcriptionist, you can obviously if you have the skills and the resources to do so, learn how to tune voice recognition models, right so that they are unique to your clients.

Yeah, that one doctor who slurs everything, because he shows up drunk every day at the office, you can train a custom model for that type of doctor.

Katie Robbert 17:53

You know, I’m trying to picture a world where it’s acceptable for a medical professional, to show up drunk every you know, literally, like

Christopher Penn 18:02

I, I have seen all sorts of crazy like it.

This is a site that always baffles me when I walk into a doctor’s office and I see like nurses or doctors outside smoking, I’m like, but but but anyway.

But there are, there are going to be you know, maybe one to 5% of the people within that profession who will go on to tuning the machines to training them to be better at what they do.

But yeah, 95% of them were like, well, I guess I need a different line of work.

Now.

There are still careers where a human has to provide the majority of the work right? It is simply not cost effective.

To teach machines to do it.

For example, cleaning a hotel room you would think is is an easy, obvious repeatable task, except that people are so random, do things like glue furniture, to the ceilings, and stuff that every hotel room that you clean is effectively somewhat different.

Sometimes there’s substantial damage, sometimes there’s, you know, a pile of pizzas or something, lift a horse in the room, whatever the case may be, that’s a case where the human being is still the best, most cost efficient resource, right? You could build an atlas robot, like from Boston Dynamics to train it to clean it real, but it’s gonna do it slower.

And it’s not going to deal with anomalies nearly as well as a person, you know, the person watching goes on for heaven’s sakes, this is not the robot horse in it.

We see it.

Katie Robbert 19:27

But I think about even like a simple example of like a Roomba, you have to train it on what the layout of your house is, in order for it to effectively clean.

It’s, you know, in a very simplistic way, it’s essentially the same as training, a large learning model to do what it is you need to do.

You know, but it also makes me think about, you know, when people are worried about will, you know, AI will take my job, there’s industries that are going to take forever to get on board.

So you know, the medical industry there There’s aspects of the medical industry that are going to get right on board with AI.

But think about how long you know, electronic health record systems have been around, you know, at least decades, I mean, a decade, if not longer.

And a lot of you know, medical practices are still struggling to get them stood up.

You know, and that’s just, you know, basic data entry connected to, you know, the clinic down the street.

You know, it’s not doing anything fancy other than just like centralizing all of your health records.

That’s not sophisticated technology, at least not these days.

And a lot of medical professions still don’t have it integrated very well.

And so when you think about a more advanced technology, like artificial intelligence, I think, at least for now, professionals in that industry, they this is a good time for them to be preparing, because they have a few years, if not maybe a decade, before artificial intelligence will take their job because of how slow moving change happens in those particular industries.

I think the legal field is another good example.

Especially if it’s if it’s a law practice that’s been around for a few generations, I don’t see them jumping to suddenly introduce artificial intelligence, and get rid of that those billable hours,

Christopher Penn 21:25

right, they won’t, but they will be in danger of, you know, the startups the new generation.

And it is it is, as much as the internet was as much as a smartphone was, the current and future generations of generative AI are a moment of substantial disruption.

The field has changed in every field.

And it is incumbent upon business owners and practitioners to recognize what those changes are to anticipate this changes as much as possible.

And to try and figure out how do we take advantage of these changes to build a more sustainable business, and to take market share from competitors that are, are more slow to react, just looking even at the search engine space.

In the news this morning.

Samsung, which is the I believe, I think is the world’s largest manufacturer of mobile phones, is in serious material conversations with Microsoft to make Bing with its the chat ma mode, the default in Samsung handsets for the search engine now because they see that it is more advanced.

This wouldn’t be happening if Google had not been so slow and behind the eight ball on a technology that realistically they should have been leading with.

Katie Robbert 22:48

Google’s and I feel like that’s a whole other topic in terms of Google’s ability to respond to what’s been happening.

And I feel like when we’re talking about, you know, AI, taking jobs, even the largest tech company in the world is going to struggle with that because of how you need to integrate it, how you need to build processes, how you need to test it, like you can’t just be like, Oh, we’re an AI shop now.

I mean, you can it’s not going to go well.

And so I think in terms of jobs that AI can’t do, is that process building, it’s that evaluation of what do we even need AI can do some of that.

But you as the human still need to give that information.

So therefore, you as the human still need to do that legwork and requirements gathering before you can even tell AI what it is that you need.

And those are the things that AI can’t replace,

Christopher Penn 23:45

and it comes down also to appetite for risk.

How much risk are you willing to take Microsoft, as the sort of the second place in the market said we are willing to shoulder disproportionate risk to take market share away from the market leader and so far it appears to be paying off.

Katie Robbert 24:04

And that’s another, you know, piece of the puzzle is AI.

AI can assess risk and give you sort of like scenario a scenario B.

But it can’t make the decision because only you will be able to know like, how much risk can I take and still sleep at night.

That’s a human decision.

That’s not a machine decision.

You know, when we we’ve pivoted our company back and forth a couple of times to be AI forward, not AI forward AI forward not AI for depending on what the market needed.

Those are not decisions that AI can necessarily make that’s more of a human task of just listening of what’s happening learning, you know, building those relationships to say like, what do you need from us and right now, our clients in the industry need us to be AI experts, and that’s fine.

We can Do that we have that skill set, but you know, six months from now, they may not.

And generative artificial intelligence will not be able to pivot the company the way that humans can.

Christopher Penn 25:09

Exactly.

So I think to sum up, AI will take a good chunk of your job, no matter what job it is, if if you as long as you have things you are doing, which are repetitive, and for now, language or image based, right or can be represented in that fashion, the three things where AI is really not capable of taking, or accountability, which is the human being has to be accountable for the majority of decisions.

Innovation, because AI can’t create anything that’s uniquely new, by its very nature, and those relationships, those human to human relationships that for most industries, most of the time people will want the exception is if your industry is so appallingly bad that people would rather deal with the machine and looking at you cable companies and the Department of Motor Vehicles.

That said, other than that, if your job does not contain a significant portion of one of those three things, a good chunk that is probably at risk.

And so Katie, what would you say? Is the read the recipe for the individual practitioner? What should they be doing right now, if they think that their career is at risk,

Katie Robbert 26:29

they should be getting a basic understanding of how artificial intelligence is going to change their particular career, because it’s not going to look the same for everyone.

And they should be looking at sort of doing an evaluation of like, okay, machines can do this, but machines can’t do this.

So let me go ahead and get skilled up on the things related to my particular career path that machines can’t do.

And you know, Chris, in your instance, if that is people management, you’re going to be up a creek, because it’s not something you’ve ever really been interested in.

And so, you know, in the example of you, your choices are to either learn how to run the machine, or learn how to manage the people.

So you’d better start learning how to run the machine,

Christopher Penn 27:16

I could do it.

Katie Robbert 27:18

No, but yeah, but for me, it’s great.

Okay, I can learn how to run the machine, but I’m better suited at running the people.

So let me make sure that I can do that to the best of my ability.

Christopher Penn 27:31

And I would say I’ll will part with recent words from marketing, thought leader Seth Godin, who said, If what you’re creating is not better than what ChatGPT can create, you need to get skilled up really fast.

And the bar is going to go up year after year after year.

So you have a very limited amount of time to get skilled up.

If you have some ideas about how you think AI is going to take jobs or not take jobs and you want to share the pop over to our free slack group go to trust insights.ai/analytics For markers where you and the 3000 other human marketers are asking and answering each other’s questions every single day.

And wherever it is you watch or listen to the show.

If there’s a platform you would rather have it on instead, go to trust insights.ai/t AI podcast, you can find us on most podcast platforms.

And while you’re there, leave us a rating or review on the platform of your choice.

It does help to share the show.

Thanks for tuning in, and we’ll talk to you next time.

Need help with your marketing AI and analytics?