Issue #3 - I will edit and humanize your AI content

March 10, 2023
Splendid Edition
In This Issue

  • “Prompt Engineer” is now a profession, and it pays more than yours
  • “Fact-checker” is now a profession, too, but it pays absolutely nothing. Discuss.
  • One-third of business leaders interviewed (in a questionable survey) say that the adoption of ChatGPT will definitely lead to layoffs by end of 2023
  • Consensus is that people mastering AI will displace people not mastering AI
  • Which leads to students having a crisis about what to study next and if it’s worth it
  • Even famous developers have their spider senses tingling
  • Meanwhile, recruiters have started a despicable thing called automated video interviews
  • So, for your next job, you might end up being interviewed by Morgan Freeman (it’s not a good thing)
Intro

Not every Splendid Edition will be an in-depth review of what’s happening in a specific industry. Sometimes, we’ll have to look at what’s happening across them to gain the right perspective.

This week, it’s time to take a look at the job market.

If you know somebody that is looking for a job or somebody that is hiring, you might want to forward this newsletter to them.

Or not. You paid for it. Be selfish. It’s your precious.

Alessandro

What's AI Doing for Companies Like Mine?

This is where we take a deeper look at how artificial intelligence is impacting the way we work across different industries: Education, Health Care, Finance, Legal, Manufacturing, Media & Entertainment, Retail, Tech, etc.

What we talk about here is not about what it could be, but about what is happening today.

Every organization adopting AI that is mentioned in this section is recorded in the AI Adoption Tracker.

When I think about how AI is impacting the job market, four different perspectives come to mind:

  • What new types of jobs are being created, if any
  • What existing types of jobs are being eliminated, if any
  • What’s happening to students preparing to enter the job market
  • What’s happening to recruiters

Let’s start with:

The new types of jobs being created, if any

The ongoing planetary enthusiasm for AI, and more specifically generative AI, is giving new job opportunities in one of the most depressed economies we have seen in a decade.

In Synthetic Work Issue #1 (Free Edition), we saw the Silicon Valley AI startup Anthropic opening one of the first-ever positions as Prompt Engineer and Librarian. A position paid an astonishing $170,000 – $335,000 for less than two years of experience.

In Synthetic Work Issue #2 (Splendid Edition), we saw a top UK law firm, Mishcon de Reya, opening a position position as GPT Legal Prompt Engineer, for an undisclosed salary.

But these are just two examples. There are many others. At the time of writing, the Boston Children Hospital is looking for a prompt engineer, too (we’ll talk about the Health Care industry adopting AI in the near future):

Now, this job title is ill-conceived: you don’t have to be a machine learning engineer to be an exceptional prompt engineer, and many prompt wizards I know are not ML experts at all. But that’s beside the point.

It’s not just in US. In Spain, the IT consulting company Raona is looking for a prompt engineer, too:

In Poland, the advertising services agency Publicis Le Pont is looking for the same:

And prompt engineer is not the only new type of job that AI is creating.

At the end of January, the giant online marketplace for freelance services Fiverr, launched six new job categories to try and address a 1600% surge in demand for AI skills in just six months:

  • AI Applications
  • AI Artists
  • AI Content Editing
  • AI Models
  • AI Music Videos
  • Fact Checking

The fact checker is critical, as the latest large language models (LLMs) like OpenAI ChatGPT or Meta Galactica spit out a torrent of lies, made-up names, historical events, research results, etc.

Content editors that can humanize the output, like Katie here, are necessary only if you don’t know how to write a prompt (can you see a theme here?). And anyway, only as long as AI systems continue to show psychopathic tendencies (I’m looking at you, Bing).

And then, of course, you want somebody that makes a music video for you. Because even if you work as an accountant, you secretly believe that you are better than Eminem, and this is your chance to show the world.

Eventually, these early specializations will go away, and the related skills will simply become part of the skillset that any copy editor, designer, producer, etc. will have to have.

In fact, if the global adoption of generative AI continues at the pace we have seen since August 2022, and if we start to see generative AI becoming part of every productivity tool out there, writing basic prompts for AI will become a skill that most Internet citizen will have to learn. Just like we learned how to do basic Google searches.

At least until newer and more capable AI models will no more require prompt engineering at all. Emad Mostaque, CEO of Stability AI (the company that nurtured the development and release of Stable Diffusion), wrote earlier this month:

Even the complex job of fine-tuning (a fancy word that means “teaching”) AI models how to reproduce proprietary content, like the design style of Pixar movies, or the writing style of Ernest Hemingway, will go away. And that is because the upcoming wave of AI models will be able to learn all these patterns from just 1-2 examples that anybody can upload online from the comfort of their sofa. Something that in technical jargon is called zero-shot learning.

We’ll see here-to-stay specializations and jobs created by AI, but they will be more advanced than anything existing today.

Next up:

The existing types of jobs being eliminated, if any

This is a very nuanced and complex conversation, and to understand in detail what’s happening on this front, we’ll have to deep dive into each and every industry, and go back and re-check the situation every 6-12 months. So, basically, you have to read Synthetic Work.

But for now, at a macroscopic level, we can already say a few things.

The first is that there’s a wide gap between what’s really happening in the job market and what the press reports about it and we must always be vigilant.

You see, reporters and journalists get paid as long as the news outlet makes money with ads (not like me, living off your generous subscriptions). And ads money comes with so-called impressions (eyeballs looking at a webpage).

And nothing like cataclysms, pestilences, economic collapses, wars, and other fear-inducing events attract eyeballs. Well, there are celebrities without make-up. OK, fine.

So it’s more rewarding and profitable for the press to paint the picture of gazillions of jobs lost to AI.

And it’s not just something people like to say about AI. As I often remember, in my last nine years in Red Hat, I spent more time explaining to the press that automation doesn’t steal human jobs than answering any other question.

Probably the most famous venture capitalist of all times, Marc Andreessen, reminds us that this question has been asked way before I started answering it:

With this in mind, let’s take a look at this statistic number reported by a Seattle-based company called Resume Builder, which polled 1,000 U.S. business leaders in February:

When asked if ChatGPT will lead to any workers being laid off by the end of 2023, 33% of business leaders say ‘definitely,’ while 26% say ‘probably.’

Within 5 years, 63% of business leaders say ChatGPT will ‘definitely’ (32%) or ‘probably’ (31%) lead to workers being laid off.

For business leaders whose companies haven’t started using ChatGPT but plan to, fewer think layoffs will result. Only 9% say the company’s use of ChatGPT will ‘definitely’ lead to workers being laid off, while 19% say ‘probably.’

48% of companies have replaced workers with ChatGPT since it became available in November of last year.

Before you start freaking out, let’s take a look at the methodology of this highly dubious poll:

respondents had to meet demographic criteria, including age (25+), income (50k+), number of employees (2+), employment status, and organizational role.

So, some of the business leaders that claimed to have replaced workers with ChatGPT might be a 25yo small business owner that makes $50K and has 1 employee.

And given that no demographic data has been shared about this attention-grabbing statistic, it’s very unlikely to be true.

So why am I wasting your time with this?

Because this news has been republished, almost entirely, by the world-famous Fortune. Which means that now there are a number of large business owners calling top consulting firms like McKinsey, Bain, or the BCG, to see if they can slash their costs by 30% by replacing employees with ChatGPT.

And guess who Bain has partnered with recently? Exactly: OpenAI.

The most likely outcome is that AI, and especially generative AI, will displace some existing jobs as it always happens when a revolutionary technology gets adopted. But the people that will see their jobs at risk are the ones that are not quick at learning and mastering the new AI tools that are emerging today.

In the words of, again, Emad Mostaque:

People that will learn quickly and continually, will have a great opportunity for profit and success.

Unless we invent the so-called Artificial General Intelligence (AGI). An AI so advanced that is capable of logical reasoning, creative thinking, and adaptation as good as a human being, or better.

In that case, I don’t know what is going to happen. Nobody does. But Synthetic Work is not focused on speculating about the future, so let’s move on.

Time to talk about:

What’s happening to students preparing to enter the job market

Synthetic Work Issue #1 (Splendid Edition) was fully dedicated to how AI is disrupting the Education industry. We saw students using ChatGPT cheating during their exams. But that doesn’t mean that they are not concerned about their career prospects because of AI.

On the Reddit forum dedicated to Stable Diffusion, one of the top generative AI systems that exists today, somebody recently asked:

The original post has now been deleted (and it’s entirely possible that this was not a good faith question, but an attempt to ignite a conversation and collect data), but the dozens of answers the poster received are still there and are worth a read.

This is just one example. There are many young people that are quite scared to ask but start to doubt their future.

Then, there are the people that already have a job and are quite experienced in what they do and yet are already concerned by the still-rudimentary capability of ChatGPT.

It’s the case, for example, of Tom Scott.

Tom doesn’t do anything to hide emotions about generative AI from his 5.83M subscribers on YouTube (and boy he must have made a ton in advertising money by doing so):

And the case of John Gruber, the well-known developer and Apple blogger behind Daring Fireball, who echoes Tom Scott.

Why are these testimonials important?

Because the visibility of these influencers raises the tension about AI and the uncertainty around one’s position in society.

I started Synthetic Work not because AI was trendy, but because I felt that we are living in a moment like no other in our history. If I am right, we must think and talk about it. These influencers seem to have similar feelings.

Finally, let’s take a look at:

What’s happening to recruiters

You probably already know that your resume has been reviewed by algorithms for a few years now. Millions of them get screened, ranked, and passed to a human recruiter every day.

And because of this, there’s an entire cottage industry that has been born to help you write a resume that doesn’t get discarded by AI. Just like with search engine optimization (SEO) techniques to improve the ranking of your website on Google, resumes can be optimized to reach humans.

There’s plenty of research that shows how these algorithms, almost always developed by a group of white caucasian male 20-something entitled software engineers in Silicon Valley show significant bias in their candidate selection.

It’s a shocking surprise, I know.

Yet, companies use these algorithms every day because it speeds up the job of understaffed recruiters.

And now the actual interviews, too, are being conducted by an AI.

Zahira Jaser, an Associate Professor, and Dimitra Petrakaki, a Professor of Technology and Organisation, both at the University of Sussex, write for Harvard Business Review:

Today, younger job seekers looking for their first role, placement, or internship are likely to face a bot at their first interview, not a human. And in the most extreme type of automated video interviews (AVIs), a bot asks a few predefined questions, giving the candidate a short window of time to answer them, and makes a decision about the person right then and there. We define these as AI-led interviews.

Just like for the stages of grief, these scholars have identified various degrees of automation in the interview process:

It sounds fun.

Last year, Anna Kramer, a reporter at Protocol, wrote a piece about AVIs. A quote from it:

The market is so fruitful that there are nearly endless options with similar services — among them HireVue, Modern Hire, Spark Hire, myInterview, Humanly.io, Willo and Curious Thing. Entry-level college graduates in tech, banking and even consulting almost always get funneled through these systems. In March 2021, HireVue announced that its platform had hosted more than 20 million video interviews since its inception.

You can even find companies that help prepare the candidate to ace the automated video interview! Look, I’m not joking:

Why are we talking about AVIs now?

Because generative AI is already making enormous progress, both with diffusion models (which help you generate pictures) and with large language models (which help you generate text).

The AI community is already producing talking faces with realistic-sounding synthetic voices and lips synched with realistic conversation led by systems like ChatGPT.

If we continue at this pace, in just a few months, human recruiters might be completely replaced by a so-called multi-modal AI. Multi-modal means capable of interacting with the user and generating outputs in multiple manners (text, images, speech, etc.).

Recruiting agencies might push for this technology as it allows them to scale their operations way beyond any human recruiter could ever do.

Hiring managers might push for this technology as the increased realism of the AI might put at ease the job seekers and they would get a more accurate screening of the candidate.

Wouldn’t you want to be interviewed by Morgan Freeman?