Breaking news: OpenAI CTO says “AI will kill jobs”

Last Modified: Jun 25th, 2024 - Category: News, UX Rants and Stuff
1353 words, 7 minutes estimated read time.
Cover image for OpenAI's CTO Mira Murati article about generative AI killing jobs

In an interview given to her alma mater, the Thayer School of Engineering at Dartmouth University, OpenAI’s CTO Mira Murati stated that artificial intelligence is likely to eliminate jobs in the creative field. She went further and suggested that some of these jobs should not exist in the first place, although she did not specify which jobs she was referring to.

In Mira Murati’s literal words:

“Some creative jobs maybe will go away, but maybe they shouldn’t have been there in the first place.”

Mira Murati, OpenAI CTO

The full interview can be viewed in the video below.

AI as a Replacement for Creative Work

It’s no secret that artificial intelligence will replace many jobs in the creative field, essentially taking over the workforce behind those jobs. In my opinion, this was expected at least a decade ago, as evidenced by the UX framework I created in 2014, Quantum UX. It’s true that Quantum UX (or QUX) was not intended to replace creative work with AI, but rather statistical and big data work. Nonetheless, if you read the What is XMI design? article, you can see that I was already perceiving a paradigm shift that would allow QUX to shine in all its glory.

In other words, it’s no secret that I have always been a fervent advocate of AI, long before generative AI even existed as a concept.

I say this because Mira Murati’s words are extremely concerning.

Firstly, what are the activities that should never have existed? By what authority can she define this?

Secondly, how would OpenAI and other generative AIs have content if it had not been created previously?

Questions that I doubt will ever be answered.

image of Mira Murati, OpenAI's CEO
Mira Murati, OpenAI’s CEO

Demystifying AI

Before proceeding, it’s important to clarify how generative AI works. As I mentioned in the previous section, when I started with Quantum UX, I defined that AI should be used as a statistical and probabilistic medium.

Generative AI follows this norm: they are probabilistic. That is, content generation is based on probability models. For example, imagine we want to generate the trunk of a tree, and each pixel has a number. It would be something like "if 1 is brown, rough, opaque," then "1.0001 is brown, rough, opaque," and so on to 1.n. Over this, resampling models are applied, which we might call libraries, which can have textures, styles, presets, etc.

What does this mean? That AIs cannot create absolutely anything. Furthermore, work done with AI cannot be registered, for a very simple reason: technically, it’s plagiarism.

Does that sound harsh? Perhaps it is. However, there are other viewpoints that “nuance” these statements. For example:

Original Content Generation: While it’s true that AIs generate content based on previous data and examples, and do not “create” in the human sense of original conceptualization or abstract thought, they can produce unique and new combinations of that data that can be perceived as original. The content generated can be unprecedented in its specific form, though derived from known patterns.

Intellectual Property and Plagiarism: AI, especially in creative fields, can generate works that are not direct copies of existing works, though they derive from a trained data set that may include those works. Copyright legislation and how it applies to AI are still developing and vary by jurisdiction. Moreover, there is an ongoing debate about whether AI-generated works should have copyright protection and who would hold those rights (the AI creator, the user, the AI itself, etc.).

I leave both viewpoints for the reader’s consideration. Just in case: the second one comes from an OpenAI source I can’t disclose, but it’s in their agenda, and there are many articles about Sam Altman asking for legislation on the matter.

Real Effects of AI on the Global and UK Job Market

Image depicting a jobless man made with OpenAI's generative AI
Image of a jobless man made with OpenAI’s generative AI. Nice keyboard!

Regardless of the above, it is a real and concrete fact that AI is already seriously affecting the job market. Personally, I know that in the field of SEO it is a big problem, as well as in low-level programming. I have not yet seen effects in creative jobs (probably because the quality of generative AI is not the best, and there are no sources like AI or PSD files). However, AI is advancing every day, and it wouldn’t be strange for this deficiency to be resolved very soon.

But outside of my subjective view, research done by the Wall Street Journal (note: requires subscription) shows that the number of orders on freelance work platforms like Upwork, Fiverr, or Freelancer have fallen by up to 21% year-over-year.

A study from UK’s IPPR says that 11% of tasks in the UK are currently susceptible to generative AI technologies that are already widely deployed. This figure is poised to rise significantly, potentially affecting around 59% of work hours when considering more advanced, integrated AI systems. This magnitude of impact is similar to the digitalization shifts since the 1990s and mirrors findings in a recent IMF study (Cazzaniga et al 2024).

The study recognizes that AI’s integration into the workforce unfolds in three distinct phases. Initially, AI tackles straightforward, ‘low-hanging fruit’ tasks—such as extracting data from digital formats—that require minimal process adjustments. Subsequently, the integration deepens, with AI systems enhancing their capabilities and access to execute more complex tasks, such as logistics management or educational grading (IPPR, 2023).

In my opinion, the most profound transformation could occur in the third phase, where AI integration depends on significant shifts in social norms and regulations. This phase could see AI stepping into roles traditionally held by humans, like giving medical advice or making staffing decisions.

What strikes me is the breadth of tasks vulnerable to AI. Routine cognitive tasks and organizational tasks face significant exposure, but even non-routine cognitive tasks that require analytical thinking and interpersonal skills are also at risk (IPPR, 2023).

However, larger enterprises appear more prepared to adopt AI, potentially widening their lead over smaller firms. This readiness could result in a two-speed economic scenario, where larger companies accelerate ahead due to faster AI adoption rates (UK DCMS, 2022).

The previous paragraph implies that if larger enterprises are more prepared, then the cost of AI adoption will be paid by freelancers and individual workers.

Similarly, an article from ABC News claims that “AI could eliminate nearly 8 million jobs in the UK, study shows,” which, if true, could prove to be a real hit to the UK job market as well as the whole UK economy.

Here is the text with grammar corrections and improvements for clarity and flow:

OpenAI’s Stance on Job Losses

In this article, I’ve shared my personal views along with some research and academic papers, mainly focusing on the UK job market. I did this not to bore you—since this isn’t an academic paper but a humble blog article.

However, after reviewing all the research and documentation (and believe me, I have read much more than the ones cited here as they are very easy to find), the comments by OpenAI’s CTO, Mira Murati, certainly add salt to the wound. In my opinion, her remarks show a complete lack of empathy for a real problem that will impact millions without any tangible benefits.

Because, let’s be honest: does anyone REALLY think that an AI-generated image, mostly made for fun, will significantly change the world for the better? Or that an article, created by compiling content from real people, will truly help anyone?

And again: I am not against AI. I fully support it. In fact, I use OpenAI’s ChatGPT to review the grammar in this article.

I’m just calling for a little bit of ethics. It’s about time.

References and Sources

“Gen-AI: Artificial Intelligence and the Future of Work” by Mauro Cazzaniga, Florence Jaumotte, Longji Li, Giovanni Melina, Augustus J. Panton, Carlo Pizzinelli, Emma Rockall, and Marina M. Tavares, published as an IMF Staff Discussion Note (SDN/2024/001) in January 2024. (download free PDF)

“AI Activity in UK Businesses” by Andrew Evans and Anja Heimann, published by UK DCMS in March 2022. (download free PDF)

We can improve your business!

Let us help you with the best solutions for your business.

It only takes one step, you're one click away from getting guaranteed results!

I want to improve my business NOW!