Overview:
Billionaire AI tech CEOs often speak highly of their AI tools as being highly intelligent and soon to be smarter than people. But the truth is that the intelligence in AI is entirely human. How high stress, low pay tech workers are powering the illusion.
Regardless of where you stand on the topic, it is impossible to ignore AI today. Whether you are scrolling LinkedIn or Facebook, you can’t get away from it. Articles, posts, memes, and opinion pieces about AI are everywhere.
Some look at the technology as a sign of a utopian society in the not-so-distant future. A technological wonder that will have robots doing our dishes, washing our laundry, and working our jobs while we enjoy life. Others, however, see it from a far darker view.
AI CEOs
If I had a penny for all the times I saw a social post comparing the rise of AI to the Terminator film franchise, I’d be rich. And while many tech CEOs speak of AI as a kind of digital savior, not everyone shares this view.
In fact, many who are anti-AI are calling BS. These may be artists, photographers, authors, and creatives. People from around the world whose work has been stolen to train AI. Often, with no compensation. Professionals who have dealt with this often do not share in the AI influencers’ positive outlook.
One thing that many tech-savvy anti-AI folks agree on is that AI is not intelligent. The main reason for this is that underpaid and overworked entry-level tech workers are training it to appear that way.
Humans behind the AI
But, in actuality, not only is AI not smart, sometimes, it is not even AI. Some companies have even been caught claiming they have a high-powered AI chatbot or agent, only to be caught using human beings answering questions and producing content for prompts live online.
So the question remains: if AI is not smart, what is causing the illusion that it is? Will all this human labor actually result in an intelligent AI? And, if so, are humans ultimately working only to replace themselves? In this article, we will seek to answer these questions and more.
What is Artificial Intelligence, and how is it defined?
Normally, I would not use AI to answer a question for an article. But considering the topic and the items here, I thought using Google’s AI Overview to answer this question actually makes sense, considering the context.
Here is what Google’s AI Overview Says
(which is powered by Gemini) gave me when asking the question above:
“Artificial Intelligence (AI) is the simulation of human intelligence in machines to perform tasks like problem-solving, decision-making, and learning, using algorithms, data, and computational power. There is no single, simple definition, but it generally refers to any artificial system that can solve tasks requiring human-like perception, cognition, planning, learning, communication, or physical action.”
Sounds legit, right? Well, sure, if it was actually simulated content generated by the machines. But here is the thing: that’s not how it works. Artificial Intelligence’s answers to questions are generated by humans.
Specifically, low-paid, overworked entry to mid-level tech workers. One of the largest subgroups of these workers in the artificial intelligence industry is data annotators.
What is a Data Annotator?
A data annotator is a technical professional who creates labels (annotates) for raw sets of data. These include, but are not limited to, audio, video, images, and text. It is the responsibility of the annotator to make these elements easily digestible for machine learning and artificial intelligence (AI) models.
They edit metadata, add tags, and provide deeper context to data to help AI systems learn to perform tasks and recognize patterns in a way that seems more human. They help the AI systems they are paid to train to generate more accurate (and human-like) output.
Workers
Many companies that employ data annotators and other AI tech workers claim these roles are essential for the development of AI. Basically, they are the gears behind AI assistants, chatbots, agents, and other revenue-generating AI tools.
However, the average pay of data-annotators and other AI tech workers tells a very different story indeed.
The stark divide in pay in the AI industry
Silicon Valley “AI Researchers” are often paid six-figure salaries by the top AI companies in the world today. However, most AI tech workers, on the other hand, such as annotators, make pennies in comparison.
Hiring annotation jobs
There are hundreds of data annotation jobs currently available. These are easy to find on job sites such as LinkedIn. And, various multimillion-dollar companies are offering these roles.
The jobs are usually made available as subcontractor positions and may or may not be full-time. The jobs often pay $20 per hour, at best. And considering these jobs are usually contracted out, when taxes come into play, these workers are taking home less than that.
But, annotators are still being paid significantly better than many other AI tech employees. These are people working jobs such as data labelers and content moderators.
Third-party Companies
Many of these workers are subcontracted by third-party companies who are using them to work on behalf of multibillion-dollar clients who are “going all in” on AI.
And yet, while these third-party companies are making bank, many of the workers they employ (particularly those overseas) are making as little as $1.46 per hour after taxes.
So, data-annotators and other AI worker bees are doing the vast majority of the hard work to create these AI systems. And while their work is turning their CEOs into billionaires, they themselves often struggle to make ends meet.
How Marketing Copy and Content Writers are also being used to train AI
When it comes to the savvy ways in which AI companies train their systems, it appears that no one’s intellectual property is safe. Many AI image and video generators are trained on millions of artists’ works.
These artists are rarely, if ever, asked if their work can be used for these purposes. And, thousands of artists the world over are expressing outrage for not being compensated when their work is “chosen” for AI training.
In the world of marketing copywriting, and content writing, the approach is more sophisticated. When it comes to tools like ChatGPT, many companies’ marketing departments freely give their own IP to AI for training. And, while some AI tools allow you to opt out of using their content for training, many companies simply don’t.
ChatGPT
They will use a tool like ChatGPT to “improve” the writers’ work. They then use the AI’s edited output for publishing in marketing, sales, and advertising. All this, while the AI system is being trained on everything it is fed.
In essence, the top AI companies’ C-suites are becoming billionaires, while those doing the heavy lifting are getting paid to eliminate their own jobs. To take this a step further, many companies are creating custom GPTs, which are essentially automated content writing systems that can learn how to write just like a company’s content writers.
As such, many companies are hiring content writers, working them to the bone for 6-12 months, training a custom GPT, and then laying them off once the system is trained to the company’s liking.
First hand experience
I myself was a victim of this trend, I took a content manager job for a technology company, training their custom AI tool on how to write blogs, I did this for 15 hours a day for 4 months.
I was then laid off once I had trained it properly. And, I am watching in real time, as dozens of my fellow marketing professionals are being laid off or fired for the same reasons while scrolling my LinkedIn feed.
AI: The good, the bad, and the truth
Despite these darker elements of the AI industry, the truth is, it is never going away. AI has already changed the way human beings do business, communicate, express emotions, create art, and so much more.
With the rise of these technologies, some tech gurus actually believe this could eventually lead to positives for many of the aforementioned workers. Concepts such as the implementation of a universal basic income are thought to be a future manifestation by many forward-thinking futurists.
Will AI actually lead us to a world where machines are working for us while we enjoy more leisure time with our families? Or will it lead to a cold and desolate future like that seen in the Terminator film series? Only time will tell.
Sources:
https://time.com/6147458/facebook-africa-content-moderation-employee-treatment
https://sigma.ai/what-is-data-annotation
Editor’s Disclaimer:
The views, thoughts, and opinions expressed in this article are those of the author, Nathan Dube, and do not necessarily reflect the official policy or position of Presence News or its editorial staff. All submissions are independently written and fact-checked to the best of the author’s ability prior to publication.
More at Presence:
