004 Shapeshifter – ChatGPT and Jobs. Australian UnderEmployment skyrockets? Curiosity and Resilience while adapting to AI

Let’s explore Shapeshifting. That very human talent to change and adapt with curiosity and exploration. How will we adapt to artificial intelligence, especially in a job market that suddenly looks VERY different? “GPTs on GPTS” the report from OpenAI, OpenLabs Research and the University of Pennsylvania gives very interesting insight into the jobs that will impacted – hundreds of them – from 19% exposure to 100%. Tax assessors don’t do too well. 100% replacement and 50% faster. But sometimes, we’re not interested in exploring or being curious or learning. We just want to get by.

I’ve got some bad news for those of you who just want to chill. This is not the time to do that. We’re giving birth to AI at the moment. The world is pushing and straining, and it’s going to come whether we want it or not. Do we get caught in a rut? That’s not my job. That’s not how this job should be. This isn’t what I planned for. I don’t understand this person. Why do I have to be the one to find common ground? Why do I have to work with AI?

I have a long and boring list of jobs that are impacted by AI/ChatGPT and how to pivot and adapt. Skip that part.

We look at Underemployment in Australia due to AI: a single person jobsharing with AI and 9 other people won’t be enough hours. Roy Morgan report on underemployment (currently at 17.1%) is more comprehensive than ABS unemployment (3.4%). Those numbers are about to skyrocket. I hope this episode has given you some thoughts about how to copilot with AI, how to move through this next phase with curiosity and wonderment rather than fear and trepidation, understanding that maybe your job will never be the same again. But that’s okay, because there’s a new approach, a new way of doing things that’ll be interesting and keep you engaged.

Resources below as well as my AI Is My CoPilot course (in Sydney, Brisbane, Melbourne, Canberra and Virtual) please go to the Australian Institute of Management website if interested: please don’t do my course unless it suits but please please please, do SOME education on AI. This thing is moving FAST.

Transcript of Alchemy of Innovation episode 4: Artificial intelligence, jobs in Australia and Underemployment (shapeshifting)

[00:00:00.960]
Alchemy of Innovation, Episode 4. Shapeshifting and jobs. So I’ve been thinking… About jobs. Actually, your job. Humanity and their jobs, careers, what they do for a living. And it seems to me, after some research, and I will be talking about the OpenResearch Lab and also OpenAI’s article on the governance of super intelligence. But after research and after thinking, I’m convinced we can work with AI, but we must prepare. And here’s the thing with human beings. They have two sides. One is they look at the stars and they say, Let’s go there. And they plan on it. We’re going to the moon, we’re going to Mars, we’re going to explore. They’re curious. Something the AI doesn’t have. Ai is not curious, we are. What’s up there in the sky? What’s in the building next door? What’s in the garden? What are the neighbours up to? What’s he doing now? We’re curious. We’re curious about each other. We’re curious about what we can learn. We’re curious about what happens when we die.

[00:01:44.890]
And yet on the other side, we sometimes react. That’s not my job. That’s not how we do things around here. He’s not my type. And suddenly, we are less curious. We just want to come home, get the kids fed and in bed, have a glass of wine, watch some Netflix. And on the weekends, go to the football, act as mum chauffeur or dad chauffeur. We’re not interested in exploring or being curious or learning. We just want to get by.

[00:02:35.210]
I’ve got some bad news for those of you who just want to chill. This is not the time to do that. We’re giving birth to AI at the moment. The world is pushing and straining, and it’s going to come whether we want it or not. I mentioned that we get caught in a rut. That’s not my job. That’s not how this job should be. This isn’t what I planned for. I don’t understand this person. Why do I have to be the one to find common ground? Why do I have to work with AI? He’s not my type.

[00:03:24.420]
My type used to be bad boy, playboy chaps who were passionate about their projects, saving the world, or starting up a business, or their careers, their football team, their mates. Actually, they were pretty well passionate about everything, except for me, the real fire in the belly types. It took me a while, about 10 tubs of ice cream and 50 packets of tissues, to stop crying and realise that maybe my type wasn’t my type. But we don’t see it when we’re in it. And maybe we can find something better on the other side of this. The job that you love today may be gone, just like those bad boys. Maybe the job of the future will be better. The boring part is taken away. And as we get used to creativity and responsibility and critical thinking, we may prefer the new job. So let me give you a little story about chat GPT that’s relevant to doing things the way we’ve always done them. There’s a brilliant young doctor and his name is Dr. Jarrel Seah, and he’s with HarrisonAI. He’s a radiologist at a hospital working on AI and radiology, medical AI basically. And he told us a really funny story about setting up ChatGPT and asking ChatGPT, giving it a query, and it didn’t know the answer. So it’s possible when you set up an instance with Chat GPT that if it doesn’t know the answer, it can set up a second instance and start agents or bots to query and find information and bring it back so that the original ChatGPT can start to query these bots and get the answers that it needs in order to find the answer, the main answer.

[00:05:25.480]
So ChatGPT set up a second instance, a child instance, and said, unfortunately, the same query, it had been asked itself. So not go away and find me some papers so I can figure out the answer. Go and search for some research, or go and search for a website for just exactly the same question. And it won’t surprise you that the second instance of Chat GPT said, No idea, no clue. So that one set up a third instance of Chat GPT. Kind of like calling a friend, but unfortunately asked exactly the same question to another identical instance of Chat GPT. And no surprises here. Third one didn’t know the answer either. It said, Bro, no clue, let me call a friend. And it set up the fourth instance. And this, of course, all happens in microseconds. So the next time Dr. Jarrel Seah will see it turned around, there were 50 million instances of chat GPT running, all of them going, No clue, dude. This is called a recursive loop. Well, not that specifically, but in programming, a recursive loop. In programming, there’s a term called recursive loop, and a recursive loop is where no new variables are introduced.

[00:06:51.470]
We date the same person, we take the same job, we do the same things, we drive the same way to work, we read the same books, we do things how we’ve always done them. And we’re exiting out of that now. No more recursive loops. No more repetitive instances of ChatGPT, no more “but this is how we’ve always done it”. You can insert the Einstein quote here about the definition of insanity, if you like. Was that Einstein? I think it’s attributed to him anyway. So when we go through major change, and if we look at any of the main verticals, we will have seen that they have transmorgified over the years. So as an example, from 1927 to 1948, the Hollywood studio star system, not just for stars, but for writers, directors, producers, creatives of all kind, the Hollywood studio system, provided an umbrella in the industry that people could take a five or seven year contract with and feel somewhat secure. That while they may not feel the most creative in that role, and certainly there was a lot of mediocre movies that came out during that time, but they had some security. The Paramount Court case, Howard Hughes’ response from RKO, caused that studio system to crash. It didn’t kill off the movie industry, but it definitely moved it into Independents. Creatives who now don’t just write the story but have to pitch it, have to find the funding, have to negotiate contracts, have to do all the things that the star system, the studio system, took care of before.

[00:08:42.240]
We’ve seen this again recently. Journalists working in newsrooms, worked in a studio system; that big media companies would protect them. And they would be left to do what journalists do best, to pose questions, to find answers, to interview. Now, many journalists are either working in a diminished capacity, or if they do have the same job, they’ve had to upskill data analytics, data research, what’s trending, and also on technology, recording audio, recording video, editing in Adobe, or Final Cut Pro, publishing online. Many struggle to stay even in those positions, and they’ve moved on. They’ve moved on to plan B, to plan C. Maybe they’re freelance, hoping for work, a hybrid traditional media writing, and also comms, media releases, writing on behalf of government, writing on behalf of corporates, maybe even writing for small business. Some have jumped ship completely and are now in comms working for PR agencies, dealing with their prior colleagues as a media relations person. And of course, they’re facing AI now, an AI that can write press releases, an AI that knows how to find the contact details of the right people, write an email, get a response, write back a response, and then create an article, all without human intervention. They keep shapeshifting, and they have to shapeshift more.

[00:10:25.320]
In March 2023, about two months ago, OpenAI and their OpenResearch Lab, which is funded by OpenAI. I think Sam Altman heads up both Open AI and Open Research Lab. In conjunction with the University of Pennsylvania, released a paper called GPTs on GPTs, and they evaluated the potential for change in many, many, many jobs. What was clear is that between 19% and 50% of what we would call white collar workers will have their job impacted by artificial intelligence using it directly. So directly means opening up generative AI, opening up something like chatGPT, or even Adobe Photoshop. So between 19 % and 50 % of jobs will be impacted by using chat GPT directly, chat GPT or something similar. Open up chatGPT, ask it to do something, write me a marketing plan, critique this statement, both positive and negative, roleplay my editor, and the response back will mean that the job is done faster. Not only faster, but more accurately.

[00:11:45.410]
What was interesting for me is that where GPT can be integrated into business systems directly into customer relationship management, CRMs directly into accounting packages like QuickBooks, directly into email, directly into Google Docs, there is a 100 % impact on the job. Not only that, but the AI will do it 50% faster. So now you have a coworker who’s cheaper than you are and 50% faster. But that’s not the end of the story. We need to navigate this carefully. And the papers, for instance, the May 22nd paper from OpenAI called the Governance of Superintelligence, does give three tasks for human beings to undertake in order to manage this transition. I think we need to stay positive, and I will be going through a list of things that we can do to attain our job, which will not be the same job, it’ll be a different job, but to retain employment.

[00:12:56.410]

Before I do that, I just want to clarify something because it’s already happening and it’s going to exacerbate. In Australia, we’re looking at about 3.4% unemployment rate and that’s from the ABS, which I’m going to question. But let’s just accept that for the moment. The unemployment rate from 1978 to 2023, so around 45 years, has been 6.68%. The highest in December 1992 was 11.2%. I believe this is going to get much higher. So Roy Morgan did a different survey which was not as limited as the ABS one. For example, with the ABS, they put potential workers in the Employed figures. Roy Morgan doesn’t. Roy Morgan found that there was hidden unemployment of 17.1 %. This number don’t meet the criteria of the ABS. They may not be on unemployment benefits, or they may not have looked for a job in the last, was it, two weeks or three weeks? But they do want to work. They either want some a job or they want more hours. And the potential workers said 1.8 million of them would take work if they could get it. Maybe they’ve given up of some kind. 900,000 wanting work or more work. When I looked at the breakdown of hidden unemployment, underemployed, about 15 % of those numbers are retail workers. They’re part time, but they’d like more. They’d like full-time, but they’re unable to secure it. Discouraged job seekers don’t meet any criteria. They’ve given up. They may be using the gig economy or some other means to survive, but about 5% of the population has tried and tried and tried to get more work and can’t. Redundant manufacturing workers, probably 20% of that underemployed population; they’ve already been affected by automation and have been laid off. They can’t find alternative employment. Selfemployed individuals struggle to get more work. Perhaps this is someone who usually pops in to help the accountant and then goes to the hairdresser and does their website for them and their Instagram. Maybe the local bakery needs a brochure, created some photos and some text. And AI is moving in on that underemployment. There are skilled workers in the gig economy. They won’t be affected. I’m talking about plumbers, electricians, carpenters. I can’t find a plumber at the best of times. About 20 % of the skilled workers in the gig economy do face hidden unemployment, but they won’t be impacted by AI. And then there are others, Indigenous job seekers in remote areas, people acting as at home carers because our system doesn’t support them. There’s a lot of people in hidden unemployment in Australia.

[00:16:11.360]
As AI takes over those jobs, the account system becomes smarter. The calling and getting leads system is more human. Creating marketing material, brochures, email newsletters are all automated. That hidden unemployment is no longer going to be hidden. These people are going to be hurting. We will see our unemployment figures go through the roof and it’s going to happen fast. IBM has already identified 7,800 jobs to be replaced by AI. It looks as though Dropbox is going to lay off about 16% of its workforce, around 500 users. Google, Apple, Facebook. Now, many layoffs will be hidden. It’s not due to AI, it’s due to the economy. Something like that. We’ll see.

[00:17:09.600]
So what are we going to do? The first thing for us to do is to get into that mindset of curiosity, of exploration. Approach AI as a tool that’s going to take all the boring jobs off of you and leave you with all the fun stuff, the creative stuff. If you have that attitude, the next few steps are going to be easier. As an example, if you’re a journalist, deepen your understanding of AI, even if it’s in a sector that’s not technical. So if you review movies, talk about AI, what are the trends, what are the ethical issues, how is it impacting scriptwriters? Make sure you know when heads roll, and they will, those journalists, writers, editors with a good grasp of AI will at least be last on the list.

[00:18:06.730]
Software developer. Yes, chatGPT can program in Python, but it needs a software developer (programmer) to read it, understand it, work with it. Make sure that you understand writing scripts and using libraries that interface with machine learning models. I would also, if I was in IT&T, be looking at CortexLabs. That’s the marketplace for models, LLMs, and so on. SingularityNet, that’s a marketplace for algorithms and data sets for AI. Become familiar with Deepbrain Chain, Huawei, and companies like that are using it to outsource infrastructure around AI. And Effective.Ai, which is a marketplace for developers. Some think like a marketplace for Facebook developers or Apple iPhone app developers. If you can bring data analyst roles into your job description, maybe start haranguing HR now, you’ll survive. Well, if anybody survives, you will. So specialise in AI related fields. If you can learn to analyse and interpret complex data sets, you’ll do better than those that have not brought data analysis into their job.

[00:19:29.000]
Remember the journalist? That can analyse trending topics, what’s important to the community. For marketing, well, you know, write better, create better. Use AI driven analytic tools to interpret consumer behaviour. Look at the behavior data and the market trends with that very human experienced eye. Engineers stay up to date with AI tools to practice critical thinking, find optimal solutions for public relations staff, comms staff, media liaison. Develop communication skills, explain an organization’s AI strategy to the public and the media in a comprehensible way. And I would add to that, make sure you are front and centre of the discussions internally within the enterprise. You be the person that explains the AI strategy to the CEO in the C-suite.

[00:20:30.760]
For lawyers, understanding AI ethics and privacy issues so that you can counsel your clients, I think is a given. For the human resources manager, engage in lifelong learning. Keep up to date with AI tools for recruiting and employee management so that you can improve efficiency. You are in a tough area. Ai is going to impact human resources in a massive way. Healthcare professionals looking at interdisciplinary knowledge, understanding how AI tools like image recognition and radiology can be integrated into patient care and recognise that doctors will be sitting with chatGPT open on their iPad as a backup health care professional.

[00:21:14.210]
Research scientists, you can network with AI professionals. You can stay up to date with their research techniquesand copilot with AI as a professional in your field. If you’re an architect, I would focus on the creativity and innovation. Using AI enhanced design software while also applying that very human creativity to create unique and functional designs will be in demand. In fact, for everybody on this list, be unique, be your brand. Do not disappear in amongst the crowd. Stand up, be counted. Project managers? Oversee a team, developing an AI application if you can. Learn to manage AI projects. Any project that does not have AI as part of its component, when money starts to run out, they will be the first ones that will be struck from the list. You want to be on the projects that are going to go ahead.

[00:22:10.910]
Financial analyst. This is the one role identified in GPTs on GPTs as being the most likely to be affected. I asked chatGPT, how will AI impact financial analysts? And chat GPT said, We can bring up real time advice based on a broad data set. If you haven’t heard me speak about big data before, basically it’s four Vs. Volume, so much data that it’s too much data. Veracity, there’s so much data that you can tell the truth in the data and leave the outliers to be liars. Variety, you can bring in data from so many different places that it starts to build the big picture. Velocity, velocity is the speed of the data. If you think about the fact that there are 6,000 tweets per second, and then you try to analyse what that means per minute, per hour, per day, by the time you’ve come up with those numbers, the AI, the algorithm, the data set, have already analysed all of those tweets, done a sentiment ranking, trending topics, big picture, small picture. But as a financial analyst, if you understand the regulatory environments, you can advise human to human those organisations that are concerned about whether they are complying with AI in financial transactions and reporting. Always remembering that when an organisation implements AI in the financial space within their organisation, they will do something called Alignment, which is to make sure that the AI self regulates and does not allow anyone to step outside the regulations.

[00:23:52.340]
Therapist and counsellor? Invest in emotional intelligence, you do already. But while AI can provide diagnostic tools and options, it’s the human’s emotional intelligence that is still critical for understanding and helping patients? Can we have online therapy with lifelike avatars and human voices? Of course. But that may not suit everybody, and you want to come through this, enjoying you are the one chosen to move into a face to face or human to human role.

[00:24:20.940]
And lastly, for fun, university lecturer, bring in AI to everything that you teach. English literature,

science, finance, doesn’t matter. Keep students updated on the latest advancements. Keep yourself relevant and topical. But I will finish on this final point because I don’t know that I made it clear. I think there’s a lot of focus at the moment on where we can see generative AI. It’s obvious. But very soon, AI is going to sink again. It’s going to sink below the waves, below the Earth. Oh, it’s still there. You’ll open up QuickBooks and it will say to you, I just did a product analysis. I think you should drop this product. There’s going to be supply issues in six months’ time. We need to plan for that now. Or you’ll open up your CRM, Customer Relationship Management, and it’ll say, “Call John. He’s going to need your product soon”. Everything’s in a blackbox. We already know that we don’t know how AI is making its decisions, and that’s going to get deeper and more hidden. The paper GPTs on GPTs calls this GPT powered. It’s where it’s embedded, it’s hidden. You take a photo with your phone and your phone says, Would you like me to add your deceased father into the photo of the family so that you feel as though he’s still there? Nobody’s going to stop and look at it and say, Oh, that’s generative AI. They’re just going to say, I miss my dad. And I do.

[00:26:01.950]
I’m going to leave it there. I hope it’s given you some thoughts about how to copilot with AI, how to move through this next phase with curiosity and wonderment rather than fear and trepidation, understanding that maybe your job will never be the same again. But that’s okay, because there’s a new approach, a new way of doing things that’ll be interesting and keep you engaged.

[00:26:31.340]
For anybody who’s interested, I have licensed my AI Is My copilot Course to the Australian Institute of Management. So if you’d like to come and do that course with me, you can. I’m teaching it in Sydney at the moment. I’m also in Kuala Lumpur in a couple of months, running a two day workshop, and in Singapore in September for a day. I think I’m there for two days actually. So let me know if you’re going to be around. I’m not saying that so you do my workshop. I think you should do some education. Find some online courses, watch some videos, something, stay up to date. This is not going away. And that hidden unemployment figure is going to skyrocket. And I do not want you to be one of those numbers. My name is Laurel Papworth. This is the Alchemy of Innovation. And remember, stay human.

List of resources

• GPTs are GPTs: An Early Look at the Labor Market Impact Potential of Large Language Models https://arxiv.org/abs/2303.10130 (Economics)
• Australian Bureau of Statistics Labour Force (April 2023) https://www.abs.gov.au/statistics/labour/employment-and-unemployment/labour-force-australia/apr-2023
• Roy Morgan (Aug 2022) Hidden Unemployed https://www.roymorgan.com/findings/the-hidden-unemployed-must-not-be-ignored-at-the-jobs-skills-summit
• OpenAI and the governance of superintelligence https://openai.com/blog/governance-of-superintelligence
• My AI Is My CoPilot course at Australian Institute of Management https://www.aim.com.au/technology/courses/ai-is-my-copilot

More alchemy of Innovation episodes.