?Where were you when ChatGPT started to disrupt your job?”

(This article is the beginning of a series of insights on how AI services are currently changing jobs, skill demands, business models and the workforce in general. It's aimed to be easily understandable and with the HR community as core audience. Please let me know if it works as intended.)

Introduction to the current situation and to ChatGPT

We all have deeply rooted memories of historical events that left a massive impression on us. In the logic of the question above, depending on your age, you might ask someone in the small talk on the landing of Apollo 11 on the moon, or the breakup of the boyband “Take That”.

During the last weeks, a seismic shift has taken place, and not everybody seems to notice the full aspects of how fundamentally this will shake the core of our daily work, our jobs and our educational system. I’m using superlatives on purpose here.

I think that a short series on the recent advent of AI tools and their influence on the workforce might be beneficial to the HR community. Currently a lot of clients seem to be worried, while others seem to be unaffected, like the day before November 30th 2022. The day ChatGPT came out.

Some of you might not heard about ChatGPT, so allow for a short description what it can do: The interface is similar to a Chat you might have with a person and the main application for most people is that you can ask questions and get answers. It is like google, with the difference that you'll get a summarized answer vs. just a list of sites you need to dig through to find an answer. You can ask it to elaborate further on details from the results it provided by asking it, e.g. “elaborate on item 3 in the result list”.

It can also e.g. summarize large amounts of text you can post into, analyze data, write letters, solve complex mathematical problems, write whole essays on a topic and it can write programming code. Real code you can instantly use. The possibilities are endless and seem like magic.

But more on that in the next article in this series.

Purpose of this series of articles

In this series I would like to deep dive into the individual AI powered tools and give you an outlook on how this will influence certain industries, jobs, skills and where you might need to rethink your workforce to stay competitive. Fast.

I’ll try to keep it entertaining and positive, despite the somewhat dystopian outlook. But no time to cry, we need to find solutions how to deal with the situation. I hope I can contribute to that by combining my knowledge of AI solutions and my experience of years of strategic workforce planning in various industries, company sizes and across the globe. Let’s engage in an open discussion, I’m happy to learn other perspectives. It’s already a vast and massively progressive topic and we need to combine knowledge and start analyzing.

“Why is this happening?” and “Why is it happening now?”

Before I start to give you some insight on individual AI solutions in the next parts of this series, let’s start with the “why”. In client discussions I always hear “Why” or “Why now?”. Depending on your job being affected, the last one might be voiced with quite some emotion.

I remember a time in 2017 when Kai Anderson and me did some talks about Digitalization, with an outlook towards how AI will influence the workforce and how it will threaten certain jobs and “DeepDream” was one example. We used a quote from Stephen Hawking in the beginning: “The rise of artificial intelligence is likely to extend job destruction deep into the middle classes, with only the most caring, creative or supervisory roles remaining.”

He was right and I would like to put the emphasis on “most”, as we see now that first and foremost creative jobs are affected by new AI tools/ platforms. If you are a Illustrator, Designer, Product Photographer, 3D artists or anybody working with text you might need to start learning today how your world will not be the same next year this time and what to do.

In our talk we had some pretty drastic examples on recent AI advancements and their influence on the workforce. I’ll post links to them and to other articles/ information below, for you to deep dive if you want to.

But let’s look at a little chronology. I’ll focus on some main events in the last years you might have noticed as well:

In the public perception it probably all started with a dream - Google “DeepDream” - a software by google employee Alexander Mordvintsev, issued to the public in 2015.

After a first awe, “DeepDream” was perceived with mixed feelings, as the images it created (“dreamed up”) tended to be on the rather creepy and dystopian side. And pretty bad in terms of image quality.

For the next years it seemed to be rather quiet in terms of large public announcement and AI seemed to remain a mainly academic topic.

But behind the curtains a lot happened and especially in the B2B business, new solutions and platforms gained traction. Something like “Synthesia.io” comes to mind, hitting the market in 2018 with their “text-to-video” generator software (see example below).

The ability to train an AI is depending on the availability of a standardized data set on a given topic. You can compare it to the same influence the availability of great learning material and teachers have on the education of a human.

A lot of progress was made here and this helped tremendously with the development of new AI solutions.

In September 2020, the Guardian released an article, entirely written by the GPT-3 language model (released as beta in June 2020) which also powers ChatGPT today. The article showed a first glimpse of what is possible, but also received some backlash (see links below).

But it brought GPT-3 to the attention of a wider public, even if access to it was not possible outside of the science community. Which, by the way, seemed to be rather concerned, as a paper from 31 OpenAI researchers and engineers shows in May 2020.

OpenAI is the company behind GPT-3 and ChatGPT, founded in 2015. In the group of founders you see names like Elon Musk (resigned from the board in 2018) and Peter Thiel. The name of OpenAi is misleading a little bit – “there is nothing open about it” as someone mentioned in a posting I’ve read a few weeks about. It’s a regular company, which just received a 10 billion investment by Microsoft in January 23.

In July 22 OpenAI issued DALL-E, a “text-to-image” generator which can produce images in a wide variety of styles – from scribbles to absolutely realistic images you wont be able to distinguish from a real photo. But in the first version it rather created a hype as a generator of weird graphics serving as foundations for bad memes.

In November 22, ChatGPT came out and OpenAI allowed other companies to include DALL-E in their applications/ solutions. ?

What we have seen in the last months is a burst of new platforms, solutions, tools and it’s difficult to keep up with the pace. One example: While DALL-E was the hype a few months ago, it is now Midjourney everyone is using for creating stock images, artistic image album covers or brainstorm industrial design or art ideas.

To answer the “Why”: It’s because there is great profit in these applications that usually come with a small monthly Netflix-like fee. And the value they can create is insane. “Paying 30$ a month to an AI for doing my job as an Illustrator?” Yes please.

The “Why now?” is greatly summarized by an article Yann LeCun (Chief AI Scientist for Meta?( Facebook) AI Research): Because OpenAI offered their solutions to the public first. LeCun seems a little bit insulted by this and belittles ChatGPT as nothing new, something others have shown in the past or working on totally similar solutions. But OpenAi certainly started this new rush by going public with their solutions.

Conclusion

The race will still intensify and companies will rush to the market to participate in these early golden days.

And based on that, the title question from this article might change pretty quickly: So please sit down for a moment and take a mental note on the question: “Where were you when Google issued their ChatGPT rival “Bard”?” (Hint: They released it yesterday)

Outlook: In the next article I’ll start with analyzing ChatGPT and it’s major influence on the workforce.

Disclaimer:

• The title image for this article were created with MidJourney, using the command “/imagine "Where were you when ChatGPT started to disrupt your job?". The creation of the image you see, took me 1 minute. Good example also of the happy mood these image generation tools tend to show ;)
• No AI was used in the creation or research for this article.

?

Link to deep-dive

Some examples from the talks Kai and me held in 2017:

AI Beats Elite Fighter Pilots, Can Run on a Raspberry Pi (designnews.com)

RBS to cut 550 jobs in preparation for robot revolution (finextra.com)

Stunning example from Synthesia.io (text-to-video creation) Interactive eLearning: Using Digital Humans | Webinar (synthesia.io)

The Guardian article written by GPT-3 in 2020: A robot wrote this entire article. Are you scared yet, human? | GPT-3 | The Guardian

Open AI Wikipedia Entry: OpenAI - Wikipedia

Dall-E showcases (all images created from pure text input only, no photoshop or other tools involved): DALL·E by OpenAI (@openaidalle) ? Instagram-Fotos und -Videos

MidJourney showcases (all images created from pure text input only, no photoshop or other tools involved): Midjourney Showcase

Article from Yann LeCun: ChatGPT resembles a slice of the human brain. That’s exactly why it’s not very smart. - The Atlantic

Article from OpenAI Researchers and Engineers on GPT-3 and the danger it can pose: ?1457c0d6bfcb4967418bfb8ac142f64a-Paper.pdf (neurips.cc)