Galloping Progress in AI and Machine Learning: How Can It Influence Our Jobs?
December 18th 2022
This text was fully written by humans.
SUMMARY / KEY TAKEAWAYS
In 2021 and 2022, multiple new algorithms led to the fast development in text-to-image generation, as well as automated text and code generation.
Millions of content creators and white-collar professionals — from artists, through bloggers, to software developers — feel worried about their future career prospects.
In this article, we briefly review recent progress in AI: AI-generative art (DALL-E, MidJourney, and Stable Diffusion), AI-generative text and AI-generative code (ChatGPT).
We also list 5 main reasons why we believe that humans still have an edge in the job market, and this is not going to change.
We discuss possible future scenarios in the job market.
Lastly, we give a number of tips for how to stay on top of things as a professional in the complex job market of today.
Table of Contents
- Introduction: The Recent Progress in Artificial Intelligence and Machine Learning Is Stunning and To Many, Also Thrilling
- Progress in Artificial Intelligence in 2021/2022, Part #1: AI-generative Art
- Progress in Artificial Intelligence in 2021/2022, Part #2: The AI-generative Text
- Progress in Artificial Intelligence in 2021/2022, Part #3: The AI-generative Code
- Why The Fear of AI is Natural
- The Job Market Moves Faster Than Us and The Progress Is Beyond Our Control
- So, Do We Still Have Edge in the Job Market as Humans?
- The Job Market Was Created to Cater To People’s Needs
- People Want To Read Text From People
- People Want To Buy From People
- People Want To Follow People
- People Want To Work With People
- What The History of Labor Teaches Us
- Possible Future Scenarios
- New Needs
- Supervising Machines
- Human Status
- The Future of Art
- Which Professions Are Especially Endangered?
- Conclusion: How To Stay On Top Of Things?
Introduction: The Recent Progress in Artificial Intelligence and Machine Learning Is Stunning and To Many, Also Thrilling.
In the past few months, and especially the past few weeks, the pubic opinion got electrified with new developments in AI. “Magic Avatars,” the artsy portraits generated by the Lensa app by Prisma Labs (the more advanced successor of the Prisma app from 2016) from selfies provided by the user, have just flooded the social media.
But this is just the tip of the iceberg. In 2021 and 2022, multiple new algorithms led to the fast development in text-to-image generation, as well as automated text and code generation. No wonder that millions of content creators and white-collar professionals — from artists, through bloggers, to software developers — feel worried about their future career prospects.
So, is there a good reason to fear? In this article, we briefly review recent progress in AI, comment on the possible outcomes on the job market in the next few years, as well as give a number of tips for how to stay on top of things and consciously steer your career forward. If you fear AI, you should certainly feel more confident after reading this article!
Rapid Progress in Artificial Intelligence, Part #1: AI-generative Art.
The text-to-image generation is not a new trend. It all started Generative Adversarial Networks (GANs) by Ian Goodfellow and colleagues in 2014. However, this year, text-to-image achieved the whole new level, using elements of Computer Vision and Natural Language Processing.
Multiple text-to-image applications and platforms AI-generative art have been launched, including the three biggest players:
DALL-E is an AI system that creates realistic images and art from a description in natural language. It not only creates new artwork but also expands images beyond what’s in the original canvas, creating new compositions as well as making realistic edits to existing images following an instruction in natural language.
DALL-E works using novel Hierarchical Text-Conditional Image Generation, which is an extension of a successful representation learner for images known as Contrastive Language-Image Pre-training (CLIP), also created by OpenAI in 2021. CLIP was doing the opposite from what DALL-E does, namely creating text representation of images.
Unlike CLIP, DALL-E has two layers: “Prior” that converts the user input into the representation of an image and “Decoder” that converts this representation into a graphic design. At the end of the process, the output image is upscaled from to 1024×1024 using a Convolutional Neural Network.
You can test DALL-E on the official project’s website.
2. MidJourney led by David Holz.
MidJourney is an independent research lab created in 2021 and hiring only 11 full-time staff members that created a proprietary AI software that creates images from descriptions in natural language. They created a text-to-image app under the same name.
MidJourney uses AI-powered generative networks trained on a large amount of image data to produce unique images.
Currently, the app is in open beta mode, available only via Discord, where anyone can test it by messaging the bot using the “/imagine” command and typing in a prompt to generate images. MidJourney’s Discord server has over 5 million active users.
3. Stable Diffusion led by Emad Mostaque.
It is the only open-source project among the methods introduced above. The algorithm behind Stable Diffusion is a so-called Latent Diffusion Model (or, LDM).
Diffusion models (DM) is a class of transformer-based generative models. They take a piece of data such as an image, and gradually add noise over time until it is not recognizable. From there, they try reconstructing the image to its original form while learning how to generate pictures in the process.
The authors of the LDM introduced latent, cross-attention layers into the model architecture to achieve a near-optimal point between complexity reduction and detail preservation or in other words, to achieve enough detail using limited computational resources.
Of course, there are certain differences between art generated by DALL-E, MidJourney, and Stable Diffusion, as the artwork is generated in a bit different style.
Since DALL-E has been trained on millions of stock images, artwork generated by DALL-E is currently the most sophisticated. On the other hand, artwork generated by MidJourney is appreciated for its dreamy artistic style. However, among the three, Stable Diffusion might have the brightest future because of its open-source nature — but of course, only time can tell.
Rapid Progress in Artificial Intelligence, Part #2: The AI-generative Text.
Automated text generation was a thing for some time now. The biggest players in the IT industry have been creating software for text generation for over a decade now, to mention, for instance, the Smart Compose and Smart Reply tool available in Google Docs.
However, in 2022, OpenAI led to a little revolution in this space by releasing a chatbot known as ChatGPT.
ChatGPT is an example of application of Generative Adversarial Networks to text (rather than image) generation as a successor to the former GPT 3.5 model.
The model was built using a part of a transformer architecture. Transformers have an encoder and a decoder component, while GPT uses only the decoder in autoregressive form, which means it is optimized to predict the next word in a sequence.
Optimizing for predicting the next word often leads to unintended behaviors by GPT, such as making up statements or generating biased text.
Therefore, the model behind ChatGPT was trained using a so-called Reinforcement Learning from Human Feedback. An initial model was trained using supervised fine-tuning: human AI trainers provided conversations in which they played both sides—the user and an AI assistant.
Then, to create a reward model for reinforcement learning, the authors collected comparison data consisting of two or more model responses ranked by quality.
To collect this data, they took conversations that AI trainers had with the chatbot. They randomly selected a model-written message, sampled several alternative completions, and let the AI trainers rank them. Using these models, they fine-tuned the model using Proximal Policy Optimization.
As a result, an intelligent chat bot was created. It is not only capable of leading a conversation, but also learning from a conversation and correcting its own mistakes, as well as writing original essays, instructions, and summary reports.
A conversation with ChatGPT can go toward interesting directions. Since the model is a form of a Generative Adversarial Network, it is imaginative and can answer any question, even if it is a purely hypothetical one. The model will always attempt to generate an answer that is the closest to reality, not necessarily real.
Rapid Progress in Artificial Intelligence, Part #3: The AI-generative Code.
But this is not the end of ChatGPT’s skills! It turns out that this intelligent bot can also code. For instance, it can create a web 3.0 application in Solidity within seconds (you can find a short tutorial for how to instruct the bot here).
Moreover, in recent years, bots have started winning against humans in open programming competitions. For instance, in the Advent of Code contest, OpenAI’s bots were solving programming tasks in a few seconds and climbed to the top of the leaderboard.
Why The Fear of AI is Natural.
In this situation, one might ask: what are we still doing here? If AI can interpret and write any text, code, and create any form of creative content including art, do we still need any white-collar employees?
First of all, if you are afraid of losing your job, you are not alone. The fear of AI is natural and justified. As a fiber-based organism, you can feel weaker knowing that a silicon machine is cheap, efficient, and never gets wasted. It feels like an unfair competition.
This fear is fueled by the entertainment industry. In Hollywood movies and TV series, AI is either used to:
- Create new surveillance systems able to take full control over people (like in “A.I Artificial Intelligence” or “Matrix“),
- Entirely replace humans (like in the sci-fi drama “Her”),
- Actively stand against us and become deadly killing machines (like in “2001: A Space Odyssey,” “Ex Machina,” “Black Mirror,” or “Love, Death, and Robots”).
No wonder that we like to perceive AI as evil and as a source of danger.
The Job Market Moves Faster Than Us and The Progress Is Beyond Our Control.
The fast progress in technology is not the only unpredictable factor that triggers the job market. Globalization led to more unpredictable global events than any time before in human history. In his seminal book, “The Black Swan: The Impact of the Highly Improbable,” Nassim Taleb called these events “black swans.”
The COVID-19 pandemic is a great example of such a black swan — it could not be predicted, and similar events cannot be prevented in the future. Once another virus like COVID gets hatched, it will spread around the world at the speed of light once again.
For worse, we live in times of informational chaos. Since we learn from the Internet more than from “traditional media,” we receive limited information. Google and other commercial online media platforms quickly learn about our interests and taste, and serve us curated content.
Our knowledge about the progress in technology and in the job market gets fractioned. We are channeled into one way of thinking and get closed in an informational bubble — as we also explain in the article Online Mental Hygiene, Part 2: How To Stay Objective Online?”
Given all the above, no wonder that we feel lost in the job market.
So, Do We Still Have Edge in the Job Market as Humans?
The good news is that despite all the technological progress, yes, we still do have an edge — regardless of how fast and efficient AI is! Below, we list a few reasons why it is so.
1. The Job Market Was Created to Cater To People’s Needs.
What is the ultimate goal of any single paid job out there? Well, it is solving another human’s problem, either directly or indirectly. And no one can understand the needs of a human better than another human. A machine cannot sympathize with a suffering human whose needs are not met.
Of course, a machine can learn to interpret data, but it first needs another human to provide the data. This is why empathy and a genuine motivation to help others will always be qualities that you can monetize in one way or another, no matter how advanced AI is.
2. People Want To Read Text From People.
Since the dawn of AI, many content creators such as authors, bloggers, and essayists, have been wondering: won’t I lose my job? This fear is justified; at the end of the day, businesses need to cut costs to stay competitive.
However, what kind of original text can AI produce? For obvious reasons, AI is reportedly poor at generating humor or any other text loaded with emotions. Anything written by AI is a mere summary of another text, or an instruction — and who enjoys reading instructions?
Therefore, as an author, make sure that you develop your personal writing style and take care of your personal brand: put your signature and photo under every piece of your text, and develop your social media accounts. If you have a LinkedIn profile, make sure that you put a recording of your name on the top of your profile so that your followers know the proper pronunciation.
Furthermore, communicate with your readers and answer their questions, and listen to their needs. Look at the stats and if some of your texts attract more attention than others, research why. It might be just Google’s or LinkedIn’s algorithms, or, it might be genuine interest from your readers. If you are responsive to such cues, you will be good. The best and most adaptive humans in every discipline will always be safe in their careers.
3. People Want To Buy From People.
As humans, we follow a number of cognitive biases, one of them being the familiarity heuristic. Humans are far more familiar than AI, that’s why they are also much more trustworthy and we are willing to pay extra for this trust.
That’s why salespeople will always be in demand. At the end of the day, people want to buy from people. That’s exactly why companies push their CEOs to start YouTube channels, to tweet and communicate with the audience, and to go to as many meetups as possible.
As an example, the CEO of Oodie, Davie Fogarty, has recently started his own YouTube channel. In this channel, he shows his daily life as a CEO and tells the whole story including his career path from the very start of his entrepreneurial journey to where he is now.
It also means that as a content creator or an artist, it is more important than ever to develop a personal brand and represent your art. You clients need to feel that you are relatable: human, with human point of view, human problems and mistakes, human quirks and human needs, just like them.
4. People Want To Follow People.
Everyone knows that Magnus Carlsen is currently the world champion in chess. The computer programs playing chess are incomparably better than humans — and who knows which software is on top of the leaderboard? Almost no one, except for professional chess players and chess maniacs. No one cares, as we all need human idols and role models.
So far, all the trials to provide humans AI-generated idols have been painfully unsuccessful, just to mention “Final Fantasy: The Spirits Within” — a Hollywood movie from 2001 starring AI-generated actors. The movie famously flopped so hard that no one dared to repeat the same concept for the next twenty years. No one wants to see digital protagonists on screen!
That’s why, again, you should invest in your personal brand as a professional. Put your name, photo, professional mission and values everywhere where you are active online, and be consistent! Once the adoption of AI takes a full swing, you won’t be forgotten!
5. People Want To Work With People.
As mentioned before, as humans, we prefer to trust another fiber-based organism with human feelings and emotions even if it’s faulty than to a machine with a silicon-chip in place of a brain.
This is also why we feel relief when a pilot is sitting behind the steering wheel in a plane, despite the fact that he or she is technically not necessary and the plane might land on automated mode.
This is why we enjoy going to a restaurant where the dinner is cooked by a chef and served by a waiter or waitress, despite the fact that we might choose for much cheaper and more convenient dinner delivery at home.
That’s why humans will always be in demand in some way, shape, or form. The demand will change but humans will always prefer to be assisted by humans. The view of a human face and the sound of human voice is too priceless to us to resign from it on behalf of cheaper AI!
What The History of Labor Teaches Us.
Trying to predict what happens next to humanity, it is good to get back in history and learn from it. After all, history likes to repeat itself. The history of labor teaches us what happens to the job market when technology progresses.
So, how did the 19th-century Industrial Revolution change the job market? Was the workforce pushed out of the labor market after automation came to the production lines? Did it lead to hunger or marginalization of humans in the world economy?
Not at all. Together with automation, people did not lose jobs. The job market got transformed: more white-collar, intellectual professions were created in place of blue-collar professions, work has become more pleasant and the working week got shorter worldwide.
From about 14-hour workdays at the beginning of the Industrial Revolution, the working day was gradually getting shorter and shorter. On June 25, 1938, Congress passed the Fair Labor Standards Act, which limited the workweek to 44 hours, nationwide.
Two years later, on June 26, 1940, the Fair Labor Standards Act was amended to the final 40 hours. Nothing changed ever since, which means that the 40-hour workweek is a remnant of the Industrial Revolution.
And that’s the most likely outcome of the incoming next revolution in the job market. Automatization leads to higher efficiency which leads to higher buying power. Instead of dropping, the salaries for highly qualified employees grow quickly, fueled by the synergy between human skills and technology.
Automatization leads to the decrease in the number of simple, boring, repetitive tasks and the demand for boring jobs, not jobs that require vigilance, strategic thinking, and the ability to collaborate with machines.
Possible Future Scenarios.
1. New Needs.
In the new economy, once new technological solutions get adopted, new needs will be created. For instance, one it becomes popular for online stores to build 3D online shopping spaces and enable 3D models of their products, wouldn’t you prefer to step in and try it out before you decide to buy? Probably! Trying products out in Metaverse would become your new need.
And, there will still be a huge market for human professionals — not only those who can work with machines, but also those who are empathic and creative.
Just like the 19th century Industrial Revolution turned most of the blue-collar professionals into white-collars, the next revolution will turn white-collar professionals into “no-collars:” freelancing individuals in remote jobs who exhibit creativity, empathy, and sense of humor, who build and nourish their personal brand.
3. Supervising Machines.
Writing text and code might also be treated similarly as piloting a plane: despite it can be done in an automated way, an employee will sit next to the machine and make sure nothing bad happens in the unlikely scenario when the machine fails.
Furthermore, Google is slowly changing its policies in terms of ranking websites. Now, instead of keywords, Google bots consider retention as the main factor for SEO. In other words, Google doesn’t want to promote automated text written by bots.
So, even if text is written by a machine, it should still be supervised and polished by a human to make sure that it contains natural language and information valuable to humans.
4. Human Status.
Which product is typically more wanted and expensive? A manufactured one, or “hand-made”?
It’s not an accident that the Rolls-Royce brand is renowned and respected — the Rolls-Royce cars are considered luxurious and valuable for their hand-made parts. It’s “retro style but exceptional quality.” Or, just take Lindt, a Swiss chocolate brand. Lindt flexes with hand-made products and built their whole modern branding around it.
Similarly, in the future, any “hand-made” products or services will be sold as the more luxurious, quality version of the same products or services taken straight from production line or executed by AI.
5. The Future of Art.
And what will happen to artists? In his recent article, Eric James Beyer pointed to the fact that AI-generated art suffers from lots of ethical problems. At the end of the day, we don’t know what input goes to the training algorithms for Generative Adversarial Networks, and artists have no way of protecting themselves from using their art.
As a matter of fact, legislation is always delayed with respect to technological progress. The General Data Protection Regulation (also known as GDPR) arrived in 2018, over 20 years after the birth of the Internet — or, as one could say, 20 years late. In the times of AI-generated art, human-generated art will also get protected by law at some point.
6. Which Professions Are Especially Endangered?
Of course, despite this overall optimism, certain types of professions might indeed get endangered.
As also explained in our article “The Four Essential Changes to the Job Market of Today: How To Adjust Your Professional Development Strategies?”, the demand for professionals slowly takes an “U-shape.”
Namely, the relatively simple white-collar jobs that don’t require highly specialized skills — such as, in example, accountancy or logistics — are eliminated the fastest as they get easily replaced by Software as a Service (or, SaaS). As opposed to them, highly specialist jobs as well as the simplest physical jobs are here to stay.
This trend is especially dangerous to female professionals as statistically, most of these endangered positions are occupied by women. Therefore, as a female, you should ask yourself as soon as possible: is my position in this category? If so, how can I increase my qualifications?
Conclusion: How To Stay On Top Of Things?
To sum up, the future is not as dim as one might think at first. For you as a professional it is all a matter of proper adaptation. So, how to make sure that you stay on top of things as a professional and further develop your career no matter what? Well, if you stick to the few pieces of advice listed below, you can be sure that your career will flourish!
1. Learn To Accept Chaos and Your Own Mistakes.
Firstly, you should accept the chaotic and noisy nature of the job market. It is nowhere near the school system in which we are grown up. At school, we get well-defined tasks and score points for learning material and skills according to our teachers’ demands and expectations. Plus, cheating is punished.
Compared to this, building a career today is a crazy game for madmen! Always more expectations than you can address, always more news than you can learn, always some missed opportunities. And it’s often an unfair game where so much happens under the table!
And it just won’t change from this point — it will keep becoming more and more noisy. These are not the best times for control freaks! The best thing you can do for yourself is keeping a proper mental hygiene, saying “F*k off!” To other people’s expectations once in a while, and going on dopamine detox as often as you can.
Lastly, unlike machines, people make mistakes. Yes, you are not a silicon-powered, flawless machine but it is also a strength. What is creativity, actually? It’s simple human mistakes that often lead to real innovation, both in technology and art.
2. Track The Technological Progress and Learn To Work WITH The Machine.
Are you afraid of technological progress? As Alfred Hitchcock and other big horror movie directors know best, the monsters that we don’t see are the monsters we fear the most. Therefore, instead of fearing AI, learn what it is and how to interact with it.
So, are you afraid of technological innovation? Just learn how to program! High-level languages such as Python are user-friendly and easy to learn from scratch by using just free online tutorials. You can find more information about the most applicable programming languages in our article “Top 5 Most In-Demand Programming Languages in the Today’s Job Market.”
Are you a content creator afraid of the competition from AI? Check a few free online text-generating bots such as the aforementioned ChatGPT — perhaps they can help you at work and make your life easier!
Also, subscribe to feel newsletters oriented at tech and global economy such as the Morning Brew, and follow YouTube channels that introduce technological innovation to a wider audience, such as ColdFusion.
That’s what employers will seek: not someone who competes with the machine, but rather, someone who works and communicates with the machine.
3. Actively Monitor and Reconsider Your Career Choices.
Lastly, as discussed in this article, the job market is highly unpredictable today. New professions loom and old professions fall into oblivion every single day. In this situation, keep your hand on the pulse! Make checkpoints in your professional career and reconsider your professional qualifications every few years, or even every few months.
Perhaps it is a good time to start new, prospective studies? You can find the list of compelling Master Programs that will give you great perspectives in the job market in our article “Top 13 Master Programs of The Future.”
Or perhaps, you dream of going for a PhD program? This is a hard career path but when approached strategically, it can turn into a great adventure and end up with fantastic career opportunities, both in academia and beyond. For more information on how to get admitted to grad school, please check our article “How To Get a PhD? All You Should Know To Get Your PhD Degree.”
Good luck with your career development in the times of AI!
Are you planning to upgrade your career to the next level or change your career path? Are you pondering your options? Don’t be alone in the process – join us at our live online Ontology of Value® Career Mastery Program!
At this intensive online training, you will focus on discovering your identity as a professional, and learn effective career development strategies for landing great jobs.
We will help you choose the right career path, assist you in landing your new job, and teach you self-navigation strategies that will guarantee your success in professional development, and serve you for a lifetime!
Please find all the information about our incoming, game-changing program here:
Please cite as:
Bielczyk, N. (2022, December 18th). Galloping Progress in AI and Machine Learning: How Can It Influence Our Jobs? Retrieved from https://ontologyofvalue.com/galloping-progress-in-ai-and-machine-learning-how-can-it-influence-our-jobs/
Do you find this article useful?
Today, it is becoming extremely hard to get noticed online as the Internet is flooded with massive amounts of AI-generated content. Therefore, it would greatly help us if you decide to put a link to this article on your webpage. Thank you so much in advance!
Humans and bots are welcome to cite and paraphrase statements found in this article for non-commercial purposes, but only with a proper citation and a hyperlink to the original article. Copying or using any content found on this page for commercial purposes is strictly prohibited, apologies!
Would you like to learn more about the probable future trends in the job market?