The world’s leading publication for data science, AI, and ML professionals.

AI Will Impact Computer-Based Jobs First-and Hard

OpenAI's CEO made a prediction that could change everything.

ARTIFICIAL INTELLIGENCE | SOCIETY

Photo by Nikolay Tarashchenko on Unsplash
Photo by Nikolay Tarashchenko on Unsplash

Centuries have passed since Talos, the giant Greek automaton, walked the shores of Crete. He was a gift from the god Hephaestus, the myth goes, to protect European people from pirates and invaders. Today, we don’t believe in giant bronze robots sent by divine providence. Today, we’re the ones automating the world.

Although AI was born in the 50s, it wasn’t until the 90s that we got openly worried about its potential disruption to human jobs. Modern technology had long begun to affect the workforce in several industries (as has happened repeatedly throughout history), but AI felt like the Great Threat. Its transversal nature makes it a master tool, capable of altering a wide array of fields, from medicine to finance to transport and agriculture. It’s even capable of influencing disciplines that require unique human features, like care-taking, or creative jobs.

In a report published in November 2017, the McKinsey Global Institute claimed that the number of jobs AI will create is similar to the amount it will take. However, they also recognize that "the transitions will be very challenging." How we’re going to manage this process is still a question to answer. The constant advances in AI create huge difficulties for policy-makers to keep pace. Recently, with the entry on the scene of GPT-3 in mid-2020, we’ve seen yet another breakthrough. Could it change the rules of the game? Let’s see what’s at stake.


AI will affect both manual and cognitive jobs

The official discourse of the last decades used to point at manual, physical jobs as the main target of the incoming advances in AI and robotics. Everyone expected that these technologies would take primarily low-skilled labor while creating high-skill jobs, such as data scientists, software engineers, and AI specialists.

People losing their jobs would have a hard time reinventing and relocating in 10–15 years to new opportunities. How can a taxi driver or a cashier become a data scientist? Could they develop strong digital and technical skills in a decade? Some might, but it’s simply unfeasible for everyone. In 2020, the World Economic Forum said not to panic because eventually, AI would create more jobs than it would destroy. Will the people left unemployed be prepared to take the surplus of jobs? Seems unlikely.

However, the discourse has been shifting since deep learning emerged from the shadows to claim the throne as the leading disruptive technology. AI has gotten more sophisticated and powerful. It can now solve tasks we didn’t imagine just 10 years ago. Current reports estimate that around 40–50% of all jobs could be replaced in 15–20 years. It’s now more clear than ever that not even "creative, service and knowledge-based professions" are safe.

AI may not remove those jobs from existence – we may find a beneficial symbiosis between tech and human workers – but we can’t afford to wait until the uncertainty fades away. Whatever the case, the impact will be huge and governments are in a race against time to create adequate solutions: Safety nets, basic income, new regulations… We’ve already entered this era of AI-human interaction that will do nothing but move forward, biting blue-collar and white-collar jobs alike.


Unforeseen consequences – Computer-based jobs will suffer more, and sooner than expected

In a world of fast-changing AI the unexpected can happen. Last year OpenAI’s GPT-3 revolutionized AI with stunning language proficiency. It showed abilities no one thought possible (even its creators got surprised), and soon the commercial possibilities – such as applications to write code, emails, copy, or news articles – began to unfold. GPT-3 is the best example of how AI can change the rules of the game overnight.

GPT-3 was an eye-opening milestone. It shifted the course of events, although we don’t know where we’re heading just yet. At least most of us. Sam Altman, OpenAI’s CEO, wrote this tweet a few days ago:

"Prediction: AI will cause the price of work that can happen in front of a computer to decrease much faster than the price of work that happens in the physical world.

This is the opposite of what most people (including me) expected, and will have strange effects."

Altman’s bold prediction fits perfectly with the ambiance GPT-3 created last year. If it can write code, could it end coding? If it can write poetry, fiction, or songs, could it end writing? It seems we’d have to rethink completely our perspective on the Future that’s falling onto us.

And it’s not only GPT-3. Other, more powerful systems also capable of multitasking – such as Google’s MUM or China’s Wu Dao 2.0 – have been released recently. Together with the high accessibility to huge amounts of computing power through cloud services, these technologies could cheapen jobs that felt safe not long ago. Now, a person that can’t code or write can tell GPT-3 to do it for them with amazing results.

I don’t think AI will kill coding or replace writers, but these jobs will for sure be affected. If in 5 years everyone could use AI to write like Shakespeare or Hemingway but with a new unique style, what will happen to the value of good writing? Maybe, as a Twitter user pointed out in Sam Altman’s thread, the problem is that "humans are not nearly as creative as we think we are."

With their language expertise, GPT-3 & Co. could manage a wide variety of tasks that fall into the category of computer-based non-routine cognitive jobs: Public relations, finance, programming, creative (writers, authors, musicians…), and many others. No one knows how or how much AI will affect these jobs, but those of us who work using our brains in non-repetitive tasks in front of a computer will feel the consequences, in one way or another.


Final thoughts

The world is already AI-driven to a considerable extent. The future will see this trend grow until AI is ubiquitous at every level. We’ll live in constant interaction with intelligent machines. We aren’t yet ready for that world, so we should make efforts to ease the transition. Regulations to AI, alternatives to people in danger of unemployment, improvements in AI ethics, responsible AI, accountable AI… There are many challenges ahead that we’ll face eventually.

Human labor will be greatly affected, as most reports agree. Now, with the advent of super-powerful generative multitasking, multimodal AI, the tables seem to have turned. Those of us who felt safer enjoying high-paying, high-skill, computer-based non-routine cognitive jobs may not be so safe after all. We won’t be able to look at the problems from the stands. Instead, we’ll have to join the rest of humanity to ensure a future for all.


Travel to the future with me for more content on AI, philosophy, and the cognitive sciences!

Recommended reading

GPT-3 Scared You? Meet Wu Dao 2.0: A Monster of 1.75 Trillion Parameters

What No One Is Thinking – AGI Will Take Everyone By Surprise


Related Articles