UX Roundup: AI Research Window | AI Skill Substitution | World Economic Forum Predicts More UX Jobs | AI-UX Consultant | Bad Bird Lip-Synch

Summary: The closing AI research window | The inflection point for AI skill substitution | World Economic Forum predicts UX job growth | Great consultant available for AI-UX work | Bad bird lip-synch

UX Roundup for January 20, 2025. (Ideogram)

Closing Research Window

We have a short window in which to conduct research on what it means when people start using AI. Right now, we can still find users who don’t use AI, and it’s also possible to study companies without AI-enhanced processes and see what happens when introducing AI. For example, many controlled studies have now been done where some people are randomly assigned to start using AI, so that we can measure the effect of AI cleanly.

Very soon, everybody will be using some kind of AI. Or, if you can hunt down a few people without AI experience, they’ll be troglodytes, introducing a strong bias in your experiment.

The research window is still open for studying the contrast between people using AI and people without AI, but not for much longer. (Midjourney)

We had a similar brief research window early in my career: in the 1980s, some workers and companies weren’t using computers yet, so we could measure the effects of computer use. Several interesting studies resulted. For example, my mentor, Dr. John D. Gould, studied business executives using word processing to write business documents and compared them with executives who created documents the old-school way: using typing pool staff to transform dictation into writing.

The executives using the word processor made many more corrections and changes to their documents than those using the typing pool, but the quality of their documents (as rated by independent experts) didn’t improve. In other words, the edits were so much busy work.

It would be impossible to conduct a similar experiment today, because no executive would know how to make good use of a typing pool.

The typing pool was a common sight in any company before personal computers. Today, no business executive would know how to make use of one. (Midjourney)

Inflection Point for AI Skill Substitution

A great example of a closing research window is the assessment of the effect of introducing AI on the job market. Dandan Qiao (National University of Singapore), Huaxia Rui (University of Rochester), and Qian Xiong (Tsinghua University) took advantage of the fact that it’s still possible to do such research and investigated the impact of AI on various freelance jobs (PDF, 10 pages).

They mainly studied two fields: translation and web development. They analyzed 350,752 freelance jobs from a major online freelancing portal. (Could have been Upwork or Fiverr, but they don’t say.) They consider the launch of ChatGPT 3.5 in November 2022 as the “introduction of AI” in these job markets. This is obviously a simplification, but they can still measure big effects. Excluding data from November and December due to holiday effects, the authors compare 5 months prior to AI with 10 months after AI.

The main findings are as follows:

• Translators: 9% fewer jobs after AI, leading to 30% lower earnings for translation freelancers.

• Web developers: 6% more jobs, leading to 67% higher earnings for web development freelancers.

This finding corresponds quite well with the skill substitution vs. complementary skills model I discussed two weeks ago:

• Translation (say, from English into Japanese) is handled so well by current large language models that AI substitutes for human translation skills. Thus, AI leads to fewer jobs and lower pay for people with those skills that have now been devalued.

• Development is boosted by current AI, but the development of a website for a small business client (a typical freelance job) requires many complementary skills, such as understanding the client’s business needs and how to translate them into computer specs. Thus, AI makes coding more efficient, increasing productivity and lowering web development prices. This leads to more jobs and higher pay for those developers who can now create more websites in less time.

Translation is a skill easily substituted by AI, meaning many human translators lose their jobs. Web development is a skill complemented by AI, making practitioners more productive and thus gaining them higher compensation. (Leonardo)

The authors ran their model on 11 different freelance jobs, estimating the following changes in freelance earnings from the introduction of AI:

• Professional & Business Writing: -32%

• Translation & Localization Services: -30%

• Information Security & Compliance: +34%

• Web Development: +67%

• Corporate & Contract Law: +67%

• Financial Planning: +78%

• Community Management & Tagging: +94%

• Project Management: +116%

• Data Mining & Management: +145%

• Photography: +348%

• Marketing, PR & Brand Strategy: +396%

It’s not clear to me why photographers earn more with AI, though maybe their “eye” is a complementary skill to image creation and video generation with AI. In general, though, these estimates also correspond well to the substitution/complementing model.

The authors introduce the concept of an inflection point for AI’s impact on a profession:

• Before the inflection point, AI has relatively poor performance in terms of the main tasks in the profession. It will help on some tasks, but humans retain the edge if they have strong complementary skills. With GPT-4 level AI, this is where we are in most of the fields listed above. At this stage, AI will lift compensation for people who take advantage of the increased productivity while leveraging their complementary skills.

• After the inflection point, AI has become good enough to substitute for most of the work done in the profession. Of the 11 jobs analyzed in this study, only business writing and translation have passed the inflection point with GPT-4 level AI performance. At this stage, AI will lower compensation for the remaining humans because they add relatively little value over unassisted AI use.

I find the author’s inflection point theory to be a compelling conceptualization for thinking about what will happen as AI moves up the value chain. We’re already expecting to see next-generation AI models this year, and more skills will certainly be substituted by those higher-qualified AI models. In particular, I think visual design and some UI design skills will start meeting the fate of writing and translation.

Most UX skills will likely not be substituted for another 3-5 years, leaving most professionals time to upskill in the complementary skills that will still be needed in the future.

While the inflection point theory is compelling, I don’t think the authors’ empirical data is nearly as compelling in proving that this is what will happen. We will have to wait for a few more AI generations before we can say for sure whether the theory is correct.

World Economic Forum Predicts More UX Jobs

The World Economic Forum (WEF) is best known for the annual Davos shindig where billionaires, prime ministers, and the occasional technology expert gather to network in the Swiss mountains. (I went one year, because they invited me and it was a famous event, but I didn’t like the sanctimonious nature of the event and haven’t been back.)

They recently released a report on “The Future of Jobs” (290-page PDF). Despite the title, this seems more of a backward-looking analysis than a forward-looking assessment of careers in the coming area of cheap and abundant intelligence provided by AI. For example, they list “analytical thinking” as the most sought-after skill. This was surely true in the past, but in the future, such activities will be performed by AI, and the human’s job will be to make sense of the AI’s findings and persuade other humans to act on them.

(Compare with my take on the skills of tomorrow: available as an article and a music video.)

I agree more with the WEF’s assessment of the next-most needed skills for the future: resilience, flexibility and agility, along with leadership and social influence. I would have rated influence (or rather manipulation) as the most important skill, since other humans will be the most important element we need to handle, when AI does all the traditional work.

Despite my criticism, there’s a wealth of interesting nuggets scattered around the report. For example, they report (based on data from Coursera) that enrollment in courses about generative AI doubled from early 2024 to late 2024. Even though we expect to see agentic AI and robotic AI become more important in 2025 and 2026, the generative AI wave has not played out yet.

Over the next 5 years, WEF expects 92 million jobs to be displaced and 170 new jobs to be created, for a net gain in employment of 7%. The jobs expected to increase the most are:

1. Big data specialists

2. FinTech engineers

3. AI and machine learning specialists

4. Software and application developers

5. Security management specialists

6. Data warehousing specialists

7. Autonomous and electric vehicle specialists

8. UI and UX designers

9. Light truck or delivery services drivers

10. Internet of Things specialists

I have seen many design influencers post triumphantly about the placement of UX design on this top-10 list, with a predicted 50% growth in UX jobs over the next 5 years. While I actually agree that UX professionals will see strong job growth in the coming decade, I don’t feel that this report is a credible argument.

Come on, listing software developers as the fourth-fasting growing job in the world? When AI improves programmer productivity so much that only explosive growth in software produced will keep their numbers steady. Or including truck drivers? When self-driving trucks are around the corner and self-driving cars are already much safer than human-driven cars. To me this list provides more proof that the WEF steers by the rearview mirror.

The World Economic Forum seems to mainly have used the rearview mirror when writing about the jobs of tomorrow. (Flux)

Despite my complaints, I do agree with that one point that UX jobs will increase. My prediction is 5x the product design jobs by 2040, or a 400% growth in 15 years. That’s pretty close to the WEF prediction of 50% growth in 5 years, which equates to 238% growth over 15 years if sustained at the same rate during the subsequent decade. (Of course, I say that we’ll be 15 M UXers in 2040, whereas WEF only predicts 10 M, so there’s still a difference. But both predictions expect aggressive growth of this profession, despite the fact that AI will be doing almost all the work that product design folks currently spend their time doing.)

Dueling predictions for the number of UX professionals in the world in 2040. I say 15 million and the World Economic Forum says 10 million. (Leonardo)

AI-UX Consultant

I was extremely pleased to learn that Ioana Teleanu has set up shop as a consultant on AI design. Her new company is called AI-R Design Studio.

Teleanu comes from a job as the Lead AI Product Designer at Miro, and while it may be unfortunate for the many UX designers who use Miro to lose her contribution to new AI features in that product, I’m sure that Miro has sufficient remaining UX staff to still do well. For the world at large, it’s great to have a consulting service available with deep expertise on how to design AI products and features.

I’m frankly disgusted with the inability of legacy UX consulting companies to provide thought leadership or good insights on AI design. There are only a handful of experts available for hire, since most AI-design expertise is currently locked up on proprietary projects. It’s great that, for example, Microsoft is hiring folks by the dozen to improve AI services for Office and Windows, and it’s wonderful that, for example, Midjourney has been inching toward better UI. But these projects build up knowledge inside those companies and don’t benefit other design teams.

Besides her job experience, Teleanu’s credentials include being the woman behind what may still be the best training course on using AI in product design. (For sure, it was the best when the first version launched in 2023.)

You can also listen to the discussion she had with me in January 2024 about AI’s impact on the design industry (Spotify, 45 min. audio). A year doesn’t sound like a long time, but in the AI field it’s an eternity, so those insights in the podcast that have held up are proof of deep vision indeed. (And if a few things didn’t turn out as we predicted, that’s the cost of being a thought leader — nobody can be 100% right, even me.)

Should you bring in an AI-UX consultant to help design your AI features or teach how to integrate AI in a new product? In the long run, I would say no: better to develop this expertise within your own design team since AI design will shortly be the only design. But unless you’ve already been AI-First for a few years, it’s worth the cost to bootstrap your AI projects with advice from one of the few people who are have done AI product design for years and are available for hire. However, unless the consultant started designing with and for AI in 2023 (or earlier!), don’t hire him or her. (Leonardo)

Bad Bird Lip-Synch

Lip synching is currently on the frontier of AI video generators’ abilities. You can see the lip-synched video I made with a human avatar a few weeks ago: decent, but not great (YouTube, 2 min. about UX trends in 2025).

However, I had a big fail when I tried to lip-synch the parrot I had used to present the findings on AI’s limited ability to interview patients in a clinical setting (Instagram, 2 min.). I guess that should be “beak-synch,” because the AI lip-synching tool simply couldn’t do it.

My expert on the recent research findings about using AI for clinical patient interviews. (Midjourney)

Compare with the original video, made with animations from Kling 1.6 without attempting to lip-synch the bird (YouTube, 2 min.). Higher image quality and just as engaging. (The YouTube version also features subtitles translated into 6 additional languages besides the parrot’s native English. But that’s a matter of which platforms support multilingual UX, not an AI issue.)

We know from Walt Disney’s pioneering work that talking and singing animals can be highly compelling. I’ll hold on to my parrot character, because it’s pretty cute and may shine in a year or so, after AI video tools have moved up a few generations.

If we learned one thing from the master of animation, Walt Disney himself, it’s that cute animals can be engaging performers. AI video tools just need to step up their game when lip-synching non-human characters. (Midjourney)

About the Author

Jakob Nielsen, Ph.D., is a usability pioneer with 42 years experience in UX and the Founder of UX Tigers. He founded the discount usability movement for fast and cheap iterative design, including heuristic evaluation and the 10 usability heuristics. He formulated the eponymous Jakob’s Law of the Internet User Experience. Named “the king of usability” by Internet Magazine, “the guru of Web page usability” by The New York Times, and “the next best thing to a true time machine” by USA Today.

Previously, Dr. Nielsen was a Sun Microsystems Distinguished Engineer and a Member of Research Staff at Bell Communications Research, the branch of Bell Labs owned by the Regional Bell Operating Companies. He is the author of 8 books, including the best-selling Designing Web Usability: The Practice of Simplicity (published in 22 languages), the foundational Usability Engineering (27,996 citations in Google Scholar), and the pioneering Hypertext and Hypermedia (published two years before the Web launched).

Dr. Nielsen holds 79 United States patents, mainly on making the Internet easier to use. He received the Lifetime Achievement Award for Human–Computer Interaction Practice from ACM SIGCHI and was named a “Titan of Human Factors” by the Human Factors and Ergonomics Society.

· Subscribe to Jakob’s newsletter to get the full text of new articles emailed to you as soon as they are published.

· Follow Jakob on LinkedIn.

· Read: article about Jakob Nielsen’s career in UX

· Watch: Jakob Nielsen’s first 41 years in UX (8 min. video)