“There are decades where nothing happens; and there are weeks where decades happen”
2023 is the year of Artificial Intelligence. Worryingly, I’m old enough to remember 2013 also being the year of Artificial Intelligence. But though history rhymes through its moral outrages and Bad Takes, it doesn’t repeat: there are clear differences between today’s algorithms and those of the last decade. Today’s most talked-about AIs are generative — far superior to their heavily supervised predecessors. So superior, in fact, that they temporarily made the tech bros forget about Web3 and crypto, which is saying something.
Generative AIs are not new. They were first theorized of in 1985, back when a neural network was still called a Boltzmann machine. But old technology has been made into new products, and that has made all the difference. Thanks to PCs, mobiles and cloud technology, anyone and everyone can now access their own personal AI. Midjourney and Dall-E make images, Riffusion makes music, Neural Radiance Fields make videos… and ChatGPT writes.
The latter is by far the most impressive. It is lighting fast and makes fewer errors than its image or sound-oriented cousins. One can now do in seconds difficult tasks that could have taken days just months ago. And so, we’ve all had the same thought upon first using chatGPT.
This is going to destroy so many jobs.
Below are the jobs that generative text-based AIs such as ChatGPT will automate, proof of that fact, and what steps to take to protect one’s career.
1. Recruitment copywriters
The danger
The recruiting industry is massive. It is massive because it is difficult to scale a deeply human experience. At some point of any classical recruiting process, someone has to understand a client brief, write a job advert, read CVs and run interviews. Through it all, a human touch is appreciated by clients, who want their needs understood, and candidates, who want to feel valued by a future employer.
But the industry is changing. We’ve known for years that assessing CVs is often automated. Now, it’s copywriting’s turn: ChatGPT can write an appealing ad while reproducing a company’s tone, if asked to.
There are 230,000 recruiters in the US today. If 15% of their time is spent crafting perfect ads, brutal internal restructuring will lead to 35,000 recruiters looking for a new job tomorrow.
Test prompt
I want you to act as a recruiting agency for developers. Write an appealing job ad for a Back-end developer. The role is in Berlin, remote 2 days a week, with competitive pay and good benefits.
How to protect your career?
Automating ads writing and CV screening is an opportunity to excel at either ends of the recruiting value chain. This means fostering better organic connections with clients to understand, refine and even anticipate their needs and expectations through unique knowledge of their situation and industry. It also means spending more time with candidates, better preparing them for interviews and coaching them at a complex moment in their career.
If ChatGPT and artificial intelligence forces us to be more human rather than less, the industry may not only survive, but thrive. Fingers crossed.
2. Software / web / front-ends developers
The danger
For years, developers have lorded their “hard skills” over the “PowerPoint monkeys.” Well, who’s laughing now?
ChatGPT can do unit testing. It can do documentation. It can provide leads to answer more complex questions, or pretend to be a Linux system, which may revolutionize bug tracking. It can even be a viable alternative to pairing, a practice lost during COVID lockdowns.
Sure, the algorithm sometimes writes nonsense, and often uses outdated packages, but this A) will improve with time, and B) also applies to most developers. Meanwhile, I don’t see it making slides.
Test prompt
Write a Python code to automatically reject a CV if it has a typo
How to protect your career?
You can become more specialised; but that will only take you so far. Coding will soon become no more than a means to an end. Which it always was.
The only way out for developers is through expending their vision beyond that of the tools made available to them. Why does something need to be coded? What does the end customer expect? Did they vocalize their need properly? Contextual knowledge can hardly be reproduced, and is the best way to beat a ChatGPT-like tool.
3. Teachers
The danger
The dangers of ChatGPT for teachers is both bottom-up and top-down. Firstly, homework in middle school and high school is now incredibly easy to reproduce via AI, and easily passes plagiarism tests. If students can use it, they will, that’s just the way of the world. English teachers now understand how math teachers felt when the calculator was invented.
At the top-down level, ChatGPT can be used to create entire curriculums, generate innovative puzzles, or pretend to be historical characters. This does not threaten teachers. What might, however, is the fact that AI tools are increasingly able to grade and correct tests. That was always the case for numerical or Multiple-Choice Questions, but now also includes longer, more complex texts.
Test prompt
I want you to act as a middle school history teacher. A student handed in the text below. Grade it and explain why you gave that grade: [hastily written summary of the French revolution]
How to protect your career?
A few methods have been floated to solve the bottom-up issue. In order to stop students from “cheating,” writing papers by hand in a classroom could work. Another more elaborate solution involves using presentations with live questions from teachers to promote speaking skills and test in-depth knowledge of a topic.
As for the top-down challenges… they’re an opportunity! If teachers can spend less time grading papers, they can use more of their time face-to-face with struggling students, which has been shown to be a net benefit for them. In fact, this is what most teachers chose their job for. Love the job, like the kids, and the rest should be OK.
4. Low-level marketers
The danger
The bottom of the marketing ladder generally promises the same to all who step onto it: create a lot of low-value content (in the form of digital ads, promotion emails, Social Media posts, blog fodder…), while minding the ever-important Search Engine Optimization.
Today, both tasks can be fully automated, and automated well by ChatGPT. This will turn the current wave of white-collar redundancy into a tsunami. Assuming there are 400,000 marketers in the US today, and that 25% of their time is spent on low value-add tasks, we’re looking at 100,000 young professionals (they always fire the kids first) needing to find a new way onto the aforementioned career ladder.
Test prompt
You are a clothing company that has mistakenly sent the wrong package to a customer. Write an apology email. Make sure to integrate an ad for one of your products at the end.
How to protect your career?
We are already too far gone to fully save the marketing industry, but there are ways to become irreplaceable within it.
Spend time creating an authentic, value-adding message that relies on the specific tone of the brand you work for. Then create a community of people who resonate with that message. No more going through the motions.
It will take time, sometimes for little results, but will ensure that one person in the company is keeping their human/writing skills sharp. We’ll need them when the robots take over and we need to convince them to keep us alive.
5. Customer Care Center employees
The danger
As we all know, Customer Care Center employees heavily rely on scripts to make or answer calls or chats. This makes them particularly exposed to automation, especially as after-sale is often seen as a major cost center.
Allow some math to come into your day: If 100 CCC employees cost $100K (incl. hiring, salary, training, and infrastructure costs) a year and do their job perfectly, the company to which they belong spends $10M a year for that perfect service. An AI can do the same job for (let’s say) a tenth of the cost: $1M a year. It only does the job well enough 95% of the time, alienating 5% of customers, who’s lifetime value is 1,000$. The company has 100K customers (1000 per CCC employee). AI mistakes are costing $5M, on top of $1M costs. The company is still up $4M, or 40% is savings (!).
It’s going to be a bloodbath.
Test prompt
I want you to pretend you are a polite customer care center employee. Write a script about a short but successful customer interaction.
How to protect your career?
More artificial intelligence is a chance for more organic intelligence, the purest form of which is empathy. The savings gained through automation of simple task should free up resources to train workers to showcase the full breadth of the human mind when faced with complex issues that require a human touch.
Employees however need to be willing to be trained, and display a level of emotional intelligence which their industry does not necessarily foster. This will need to change.
6. Journalists
The danger
The job of a journalist is not as exciting as what we see in movies. Some may topple presidents, but most will cover sports, elections, or product launches. When in doubt, assume you belong to the latter category.
AI today can automate a lot of this. Enter a few parameters into a pre-existing prompt, and the article just writes itself. Sure, AI is making errors today, but that won’t last long. ChatGPT3 already writes at the level as of a High School student happy with a B-. Who knows where ChatGPT4 (planned for a 2023 release) will take us...
Test prompt
I want you to act as an insightful technology blogger. I will give you a prompt about a new technology, and you will write an in-depth press release of over 600 words, including quotes, pros, cons, features, and comparisons to other technologies on the market. Please optimize it for SEO on Google (especially the first paragraph). The first announcement is about Netflix CEO’s resigning.
How to protect your career?
It may be tempting to leave journalism to the algorithms and concentrate on the opinion section. This would be a mistake. Opinions are a muscle, and require constant work. But there are only so many opinions to be had before they become toxic.
The real value for journalists today is in the data unavailable to ChatGPT: non-public stories. Secrets people are hiding, neighbourhoods no one goes to, trends no one has spotted… We need investigative journalists, especially local ones. Leave your desk, talk to people, and print what they have to say — that’s how you beat the algorithm.
7. List guys
The danger
Anyone who’s been on Twitter or LinkedIn for a hot minute knows about the men (and sometimes girlbosses) who identify as “founders” and spend a lot of time writing cringe threads. They often use a 🧵 emoji and write tips for making money from home, or lists of billionaires’ supposed habits.
They use threads to promote books that help you write books that help you sell courses that help you write newsletters that help you sell the books that help others write books that help them sell courses tha — . All of which can now be done with ChatGPT in less than 10 minutes. Hustlers need to up their game to survive.
Test prompt
I want you to act as a Twitter Productivity Influencer. Give me 10 ways to make money with ChatGPT. Make sure they are all written as clickbait.
How to protect your career?
I don’t think this is a “career” that should be protected. But for the sake of the exercise, here’s how it could be done: hustlers need to start spending more time creating a trusted brand. As content becomes commoditized (and thus infinite), trust of its quality through its source will matter immensely. That means influencers need to investigate a niche, understand it, and craft an original message that overshoots the material world and squarely lands in the realm of emotion, which cannot be reproduced.
If that sounds like authenticity, you’d be right. The age of “financial advisors” posing in front of rented Ferraris is over. And not a minute too soon.
We love to write the words “This will change everything.” Then we go back to our regular lives, too busy to realize that change is happening very, very slowly, if at all.
ChatGPT, just like any tool, is only as good as those who wield it. It will hurt the people who are not in roles they should be in, be they the lazy or the mediocre. Hopefully, they will find thing they are better suited for. As for the talented and passionate, they will be made infinitely better by the technology.
I only worry this leads to the creation of a fully fractured society between those able to use technology, and those replaced by it.
There are decades where nothing happen, and weeks where decades happen.
That quote is from Lenin. If you feel you understand why the added context makes this article scarier, funnier, or both… trust those instincts.
Good luck out there.
