Economics of AI in UX: Why we should be optimistic

22 May 2025 10:07 AM | Laura Cunningham (Administrator)

Author: Caleb Furlough, PhD

Advancements in AI continue to generate excitement for those in tech and product-focused fields. Coming along for the ride with that excitement is proportionate anxiety for many[1], especially those in UX. New AI advancements and employer expectations raise various concerns that tend to fall under the larger banner of something like “will AI reduce the need for UX designers and researchers?”. Thomas Stokes at Drill Bit Labs reports a year-over-year increase of almost 7% in UX job postings mentioning AI[2]. Are User Experience and related fields like Human Factors and Human-Computer Interaction in the early phases of being relegated to the sidelines? While I understand and empathize with this concern, I also believe it is often misplaced. Fundamental economic principles and recent historical patterns suggest something more nuanced and optimistic. The UX workforce is more likely to see its daily job responsibilities and overall market positioning transform and evolve than become outdated and marginalized. 

My goal for this post is simply to offer a few thoughts on the future of UX based on proven economic principles and historical examples I hope will inspire optimism.   

Foundational Economic Principles

Taking a look at recent economic history, technological advancements have consistently transformed rather than eliminated human work. I will briefly highlight two relevant economic principles to this transformation: Ricardo’s Comparative Advantage and Kremer’s O-ring Theory. 

Comparative Advantage: In 1817 renowned economist David Ricardo demonstrated that even when one side of an economic trade has an absolute advantage in everything they produce, both sides will benefit when each specializes in what they do comparatively better. In other words, even in the extreme hypothetical that an AI agent was better at every UX task than its human counterpart, it would still be more efficient to utilize the AI for only the tasks it has the largest comparative advantage in (and likewise for the human UXer). This principle helps us see how, even as AI capabilities increase, economic forces will create a division of labor between AI and humans rooted in their respective strengths (even in the unlikely, doom-and-gloom scenario in which AI has become absolutely better at everything). 

O-ring Theory: Nobel Memorial Prize winning economist Michael Kremer proposed the O-ring theory of economic development in 1993[3]. Using the Challenger space shuttle disaster of 1986 as a prime example, O-ring theory shows how a single point of inefficiency in a product team can dramatically impact the success of the end product. For example, even if 9 out of 10 team members do their job with incredible efficiency, a single team member working inefficiently can cause the otherwise high performing team to produce a mediocre product. Applied to our AI and UX context, Kremer’s theory demonstrates that even if we were to efficiently outsource 90% of UX tasks to AI, the relative impact of the human UX designer or researcher would not decrease, but actually increase!

Takeaway: Comparative Advantage and O-ring theory provide two economic rationales to believe there will always be considerable value to be gained from human UXers, even in the event humanity reaches sci-fi levels of AI capabilities. 

Historical Examples of Role Transformation

Theories and principles are helpful, but let’s turn for a quick look at some historical examples showing what happens when technology automates human tasks.  

  1. Automation and mechanization in manufacturing is a familiar example. When mechanized task automation became the standard manufacturing it did not simply eliminate jobs, but instead bifurcated the market. Yes, mass production was automated and job roles were lost as a result. However, this simultaneously created new demand for artisanal and specialty manufacturing roles. It is estimated that in the past 25 years roughly 1.7 million manufacturing jobs have been lost due to automation[4]. However, it is also estimated that 133 million new jobs (some, not all, in manufacturing) will be created, many of which will shift from low-skills and low-pay to higher-skilled and higher-paid manufacturing positions [5]

  2. The banking industry gives us one of my favorite examples of transformation over elimination. When ATMs became widespread in the 90’s many spoke of the impending doom of the bank teller. The number of ATMs sharply rose from zero to over 400,000 in the course of just a few years, bolstering fears of job loss. Instead, as documented by economist James Bessen[6], the number of bank tellers in the United States increased at a rate higher than that of the average workforce. ATMs reduced the cost of operating a bank branch, allowing banks to open more branches. Teller responsibilities transformed from handling cash to relationship management and more complex financial services. These were high-value responsibilities that ATMs could not perform at the level of a human or, in Ricardo’s language, in which humans held the comparative advantage

Bessen concludes 

“Many people suppose that if technology automates tasks . . . then widespread computer automation must be associated with major job losses. But this view fundamentally misunderstands what has been happening. The evidence shows that computer automation of an occupation is associated with increased demand for that occupation . . . The net result is that computer use is associated with a small increase in employment on average, not major job losses.” (pp.29-30)

3. When thinking of large-scale automation in North America, the agricultural revolution is one of the most dramatic historical examples. Agricultural employment fell from roughly 41% of the U.S. workforce in 1900 to less than 2% today, due in large part to technological advances in automation[5]. The end of humans working in agriculture, right? Not quite, as both The Bureau of Labor Statistics and economist David Autor note that while traditional agricultural jobs did experience decline, new jobs and responsibilities were created focusing less on physical labor and more on higher-order thinking[8],[7]. This is not to mention the significant rise of new jobs and skill sets in other sectors as a result of this market shift.

4. Lastly, a historical example more directly applicable to what we see happening is UX is the transformation of accounting. As standard bookkeeping has become increasingly automated, accountants have evolved into new and sometimes different, or more highly skilled, roles such as strategists, analytics, and advisors[7]. The Bureau of Labor Statistics reports employment among accountants has grown and is expected to continue to grow at an above average rate despite automation of many daily tasks[10]. Traditionally trained accountants are increasingly focused on higher-order responsibilities like interpretation, strategy, and personalized financial advice. It is not difficult to see how a similar strategic shift could happen in UX. 

These historical patterns suggest UX may begin to bifurcate into AI-driven standard solutions and value-add, human-crafted experiences. This would follow the economic principle of comparative advantage that has a proven track record through many technological revolutions.

The Current State of AI in UX

Today's AI tools are capable of performing, at varying degrees of efficiency, different core UX tasks like generating wireframes and prototypes, sketching design concepts, creating synthetic users, moderating 1-on-1 interview sessions, mapping user journeys from qualitative data inputs, drafting moderator guides, analyzing quantitative data, creating research reports, ideating new design directions, and the list goes on. 

However, as we stand today, AI struggles with certain aspects of UX like facing unique design challenges, cultural and contextual nuance, novel user behaviors, and ethical considerations. Jakob Nielsen comments on some of the current limitations in AI in UX Research, 

"The reason we conduct user research instead of relying solely on the 10 heuristics and other usability guidelines is that humans always have unexpected behaviors. How do we know what people want, need, or prefer? This cannot be predicted, any more than we can predict the future.” [11]

Perhaps these limitations reflect Ricardo’s principle of comparative advantage at work. AI handles routine tasks along with some creative tasks while humans retain the advantage in other areas of higher-order thinking. I try to avoid making specific predictions about the future, but  this seems like a reasonable possibility given what we have seen happen in other technological revolutions. 

Looking Forward

I will end by offering a few parting considerations for looking at AI transformation in UX moving forward. UX professionals navigating current advances in AI should consider adopting the lens of tried and true economic principles for guidance. Focus on developing skills where humans maintain comparative advantage. Do not become fixated on the absolute advantages of AI. Focus on core UX skills that make you part of the O-ring of production to multiply your impact on the end product. Focus your energy not solely on displacement, but also on newly created opportunities. 

As bank tellers and accountants embraced evolution in their roles through a changing technological environment, we UX professionals should too - with optimism. 

References

[1] https://www.reach3insights.com/blog/ai-anxiety

[2] https://substack.com/home/post/p-162155396

[3] Kremer, M. (1993). The O-ring theory of economic development. The quarterly journal of economics, 108(3), 551-575.

[4] https://teamstage.io/jobs-lost-to-automation-statistics/

[5] Jiang, H., Ge, Y., Yang, C., & Yu, H. (2024). How automated machines influence employment in manufacturing enterprises?. PloS one, 19(3)

[6] Bessen, J. E. (2016). How computer automation affects occupations: Technology, jobs, and skills. Boston Univ. school of law, law and economics research paper, (15-49).

[7] Autor, D. H. (2015). Why are there still so many jobs? The history and future of workplace automation. Journal of economic perspectives, 29(3), 3-30.

[8] https://www.bls.gov/opub/mlr/2019/beyond-bls/changes-in-the-us-occupational-mix-from-1860-to-2015.htm

[9] https://www.getcone.io/blog/the-evolution-of-an-accountants-role-modern-accountant

[10] https://www.bls.gov/ooh/business-and-financial/accountants-and-auditors.htm

[11] https://jakobnielsenphd.substack.com/p/ai-can-cannot-do-ux



Copyright © Triangle User Experience Professionals Association

Powered by Wild Apricot Membership Software