The creators of recent AI phenomenon, ChatGPT, have co-authored a research paper looking into the impact of large language models (LLMs) on the labour market. It highlights the expected impact of these generative pre-trained transformer (GPT) models on different, specific occupations, and even gives a list of the jobs it has zero impact on.
So, if you’re lucky, talented, or dedicated enough to be a professional athlete your job’s safe. Until they give the AI legs as well as language. What? You mean they’ve already got legs? Holy shit, we’re all toast.
In all there are 34 different occupations that are listed as having no exposure—that’s the paper’s metric for the impact of LLMs—but it also states that 80% of the US workforce will see some impact on their roles.
The paper, titled “GPTs are GPTs: An Early Look at the Labor Market Impact Potential of Large Language Models” (graph-filled PDF warning) is co-authored by OpenAI, OpenResearch, and the University of Pennsylvania. So, if we’re honest we’re going to have to take its findings with a modicum of salt given how much skin the authors have in the LLM game.
But if you want to be confident of resisting the AI revolution then you’re going to want to be retraining for one of these occupations:
- Agricultural Equipment Operators
- Athletes and Sports Competitors
- Automotive Glass Installers and Repairers
- Bus and Truck Mechanics and Diesel Engine Specialists
- Cement Masons and Concrete Finishers
- Cooks, Short Order
- Cutters and Trimmers, Hand
- Derrick Operators, Oil and Gas
- Dining Room and Cafeteria Attendants and Bartender Helpers
- Dishwashers
- Dredge Operators
- Electrical Power-Line Installers and Repairers
- Excavating and Loading Machine and Dragline Operators, Surface Mining
- Floor Layers, Except Carpet, Wood, and Hard Tiles
- Foundry Mold and Coremakers
- Helpers–Brickmasons, Blockmasons, Stonemasons, and Tile and Marble Setters
- Helpers–Carpenters
- Helpers–Painters, Paperhangers, Plasterers, and Stucco Masons
- Helpers–Pipelayers, Plumbers, Pipefitters, and Steamfitters
- Helpers–Roofers
- Meat, Poultry, and Fish Cutters and Trimmers
- Motorcycle Mechanics
- Paving, Surfacing, and Tamping Equipment Operators
- Pile Driver Operators
- Pourers and Casters, Metal
- Rail-Track Laying and Maintenance Equipment Operators
- Refractory Materials Repairers, Except Brickmasons
- Roof Bolters, Mining
- Roustabouts, Oil and Gas
- Slaughterers and Meat Packers
- Stonemasons
- Tapers
- Tire Repairers and Changers
- Wellhead Pumpers
They’re all pretty self explanatory, and mostly based on skilled manual labour. Though I’m a little concerned about just how much effort was put into this research as it mentions the occupation of ‘dishwashers’ is going to be fine. I mean, if we’re honest, AI hasn’t got a lot to do with replacing human dishwashers, those mechanical dishwashers are doing pretty well at that right now.
But if you don’t want to be royally screwed by future LLMs then you will need to make sure to shy away from such occupations as… checks notes… journalism. Ah.
In all, the paper suggests that 19% of jobs will see at least 50% of the work they do exposed by ChatGPT and its analogues. But that doesn’t necessarily mean replacing the work they do, but potentially augmenting or assisting it.
In fact, OpenAI CEO, Sam Altman, has gone on record speaking about using AI models as tools rather than something that can entirely replace a human counterpart.
“It waits for someone to give it an input,” Altman says in an interview with ABC. “This is a tool that is very much in human control.”
This is a view shared by Google, as I asked Bard—it’s own ChatGPT rival—should I be worried about my job as a technology journalist?
On the whole it’s response was very positive, speaking about the demand for tech journalists increasing as more people rely on technology to stay informed and connected. Then it gave me a little advice about developing sources, and being concise, and not rambling as I’m wont to do.
“Cools, thanks,” I say. “What about the impact of LLMs like you?”
“LLMs like me can have a significant impact on technology journalism. For example, I can be used to generate large amounts of content, such as news articles, blog posts, and product reviews. This can free up journalists to focus on more in-depth reporting and analysis.
Additionally, LLMs can be used to translate content into different languages, which can help to reach a wider audience. And, I can be used to generate personalized content, such as news articles that are tailored to a reader’s interests.
Overall, LLMs like me can be a valuable tool for technology journalists. However, it is important to note that I am not a replacement for human journalists. I can only do what I am trained to do, and I cannot replace the human touch that is essential to good journalism.”
Which is nice to hear. Though my main concern with LLMs has been about how confidently they can lie, or deliver incorrect information. Even about something seemingly so simple as the number of characters in a sentence. So yeah, I’m still not sure if I ought to be taking career advice off either Bard or ChatGPT just yet.
The chances of my becoming a professional athlete have long since passed, so I guess I’ll retrain as a mechanic. I don’t think I’ve got the stomach to work in an abattoir.