The Case for Licensing Product Designers
Product designers who shape high-risk digital systems should be subject to a professional accountability framework comparable to the licensing, standard-of-care obligations, liability insurance, and safety-intervention mechanisms that govern architects and engineers.
The reason is simple: software products are no longer merely tools. They are public infrastructure for communication, education, commerce, identity, civic life, and childhood development. When they fail, the harm is not always visible in concrete, steel, or smoke, but it can still be widespread, foreseeable, and profound.
I learned the practical force of professional accountability during the construction of our house in 2022. We razed an old house built in 1900, leaving only fragile CMU basement walls held together by century-old mortar. Because the new house would have a larger footprint and a heavier live-load roof, those walls needed reinforcement to meet modern load-bearing and seismic standards. Our architect and engineer both missed critical information: the architect failed to communicate clearly about the CMU walls, and the engineer assumed the foundation was concrete without inspecting it in person. The mistake caused a projected $70,000 cost overrun.
We were angry, but we also had leverage. A second architect confirmed that the team had failed to meet the appropriate professional standard of care.[1] In the end, the architect and engineer cooperated, corrected the design, and helped file an insurance claim. They did so because they were licensed professionals. Their work was subject to scrutiny by governing boards, professional standards, and liability insurance. Their reputations and livelihoods depended on accountability.
That experience influenced my current thinking about product design. In architecture and engineering, public safety is not left to corporate goodwill or individual conscience alone. It is embedded in professional duties, legal standards, inspection regimes, insurance requirements, and equal power to stop dangerous work. Digital products now shape public life with comparable consequences. Yet the people who design, approve, and deploy them operate with far fewer – if any – professional constraints.
This essay argues first that architecture and engineering offer a useful model for professional accountability; second, that software harms are now serious enough to justify a similar framework; third, that recent litigation against major technology platforms shows the inadequacy of self-regulation; and finally, that Washington state is well positioned to pioneer a licensing framework for high-risk digital product work.
Lessons from Architecture and Engineering
Professional licensing exists because some kinds of work create risks that ordinary consumers cannot evaluate on their own. A homeowner cannot independently verify every structural calculation behind a house. A patient cannot audit every clinical judgment required for a surgery. A driver cannot inspect every arch beneath a roadway or cable joint on a bridge. In these fields, society responds by requiring education, examination, ethical duties, continuing competence, liability insurance, disciplinary review, and enforceable standards of care.[2]
The Tacoma Narrows Bridge collapse remains one of the United States’ most famous engineering failures. The bridge opened in 1940 and collapsed only months later after wind-induced motion revealed design vulnerabilities that engineers did not yet fully understand.[3] The failure was spectacular, but the lesson was institutional: expert review, public scrutiny, and professional learning produced new standards for suspension bridge design.[4] Engineering changed because failure triggered accountability.[5]
Software applications are not bridges, but the point is that complex technical systems can cause public harm when designs fail and there are no clear, consistent standards to hold people making design decisions accountable. When physical infrastructure became too complex and dangerous to leave unregulated, society developed licensure, standards, inspection, and stop-work authority. Digital infrastructure has now reached a similar moment.
Three concepts matter most. Licensing creates minimum qualifications and professional discipline. A standard of care defines what a reasonably competent professional should do under comparable circumstances. Stop-work authority gives workers and regulators a mechanism to pause dangerous activity before harm escalates. Software needs all three, adapted to the realities of digital product development.
From Physical Harm to Digital Harm
For much of the software industry’s history, design mistakes were treated as usability issues or business problems of generally minor consequence. A confusing interface might frustrate users. A spreadsheet error might create low-stakes financial errors. A failed product might cost a company money. But product design was not widely understood as a profession capable of creating catastrophic public harm.
That assumption no longer holds. Digital products can influence childhood development, mental health, sleep, body image, civic discourse, consumer behavior, public safety, and democratic trust. Algorithms can amplify harmful content. Interface choices can encourage compulsive use. Recommendation systems can target vulnerable users. Defaults can nudge behavior at massive scale leading to catastrophic political outcomes and even death as we saw with Facebook in Myanmar. Most of the harms may be largely psychological, behavioral, developmental, financial, or civic rather than structural, but they are still public-safety harms when they are foreseeable, systemic, and preventable.[6]
This distinction matters because software harms are often dismissed as too intangible to regulate. Yet the law already recognizes nonphysical harms in many contexts, including consumer protection, privacy, discrimination, negligence, product liability, and professional malpractice. The question is not whether harm must be visible to matter. The question is whether professionals who create foreseeable risks should have enforceable obligations to reduce them.
Social Media Litigation and the Failure of Self-Regulation
Recent lawsuits against major social media companies show why voluntary self-regulation is inadequate.[7] Claims brought against Meta and Google have focused not only on content, but on product design: recommendation systems, engagement incentives, filters, notifications, and other features alleged to encourage compulsive use or expose children to foreseeable harm.[8] These cases matter because they frame platform harm as a product design problem, not merely a speech problem.[9]
The New Mexico case against Meta is especially meaningful to me. I spent part of my childhood in northwestern New Mexico, growing up in the high desert from fifth through tenth grade. If a case like this had existed forty-five years ago, I would have been one of the children its attorney general sought to protect. New Mexico is often ranked near the bottom of U.S. states on measures such as education, opportunity, and infrastructure,[10] but it is also a state of extraordinary beauty, Native culture, scientific expertise, and personal significance. It is the birthplace of Microsoft, where I later worked as a product designer for fourteen years.
The personal connection matters, but the policy lesson matters more. If internal warnings, public reporting, regulatory pressure, and litigation are all required before companies respond meaningfully to harms affecting children, then the system is reactive by design. A professional accountability framework would not eliminate harm, but it would create duties before litigation, not only after tragedy.[11]
In construction, a worker who sees imminent danger – regardless of title or position – can stop work until safety concerns are addressed.[12] In software development, stop work orders don’t exist. Employees may raise red flags about harms to children or vulnerable users, only to see those warnings minimized, reframed, delayed, or overruled by business priorities. Product designers and researchers often identify risks before software is released to the public, but they lack the independent professional standing, legal protection, and external reporting channels that licensed professionals in other fields possess.
Product Design and Standard of Care
A standard of care for high-risk digital product work would not require perfection. Architects are not liable every time a building has a flaw, and engineers are not expected to foresee every unknown unknown. The standard asks whether a professional acted with the skill and care ordinarily expected under similar circumstances.[13] Product design needs an equivalent benchmark.
Such a standard could address whether a product team identified foreseeable harms, consulted relevant experts, documented tradeoffs, tested vulnerable-user impacts, escalated unresolved risks, and avoided design choices known to exploit cognitive or developmental vulnerabilities. It could distinguish ordinary product iteration from reckless design. It could create a record of professional judgment when companies choose engagement, growth, or revenue over safety.
My experience designing Excel maps at Microsoft showed that careful product governance is possible. Choropleth maps require geographic, political, and cultural judgment. Map projections distort land masses. Icons can imply geopolitical preference. Country boundaries may be contested. Because Microsoft operated in a global legal and regulatory environment, we spent significant time consulting GIS experts, iterating designs, and working with legal teams. The process was slow, detailed, and sometimes frustrating, but it reflected a basic truth: when law and accountability require care, product teams respond accordingly.
The lesson is not that large corporations are incapable of responsible design. The lesson is that they often act responsibly only when external incentives require it. Self-regulation is not enough when the most profitable design choices can also be the most harmful.
A Licensing Framework for High-Risk Digital Product Work
The goal should not be to license every person who contributes to software. That would be impractical and unnecessary. Instead, licensing should apply to professionals who hold responsibility for safety-relevant design decisions in high-risk digital systems. The trigger should be function, not job title.[14]
A workable licensing framework could include the following components:
· Covered systems: Products that affect women, children, mental health, education, finance, employment, housing, public benefits, health care, civic participation, biometric identity, or other high-risk public interests.
· Covered professionals: Designers, researchers, product managers, design leaders, and other decision-makers with authority over safety-relevant user experiences, recommendation systems, persuasive design patterns, data collection practices, or vulnerable-user impacts.
· Core competencies: Human-centered design, accessibility, child safety, privacy, algorithmic accountability, behavioral risk, research ethics, inclusive design, product liability basics, documentation, and escalation duties.
· Standard of care: A legally recognizable duty to identify foreseeable harms, document tradeoffs, consult experts where needed, and avoid design patterns that create unreasonable risk.
· Continuing education: Ongoing training in emerging technologies, child development, privacy[CN1] law, AI safety, accessibility, and regulatory updates.
· Professional liability insurance: Coverage comparable in purpose to professional liability insurance for architects and engineers, scaled to role and risk.
· Whistleblower and stop-work protections: Protected channels for licensed professionals to escalate serious safety risks internally, to regulators, or to an independent board without retaliation.
· Disciplinary process: A state board or professional commission empowered to investigate violations, issue sanctions, suspend licenses, and refer matters to consumer-protection or attorney-general offices.
· Organizational accountability: Requirements that companies employing licensed professionals maintain safety documentation, risk reviews, and escalation records rather than shifting all blame to individual workers.
This framework would not turn product designers into scapegoats. Instead, licensing would give practitioners professional standing when they identify risks, and it would make it harder for companies to bury safety concerns beneath growth metrics. It would also create a clearer evidentiary record when companies ignore documented warnings.
Objections and Limits
Any licensing proposal must confront serious objections. The first is that licensing could suppress innovation.[15] That risk is real if licensing is too broad or bureaucratic. The answer is to focus on high-risk systems, not ordinary websites, small-business tools, or low-stakes consumer software. Regulation should be proportional to foreseeable public harm.
The second objection might be that software is too collaborative to assign responsibility to individuals. But architecture and engineering are also collaborative. Licensing does not eliminate organizational liability; it clarifies professional duties within complex systems. A digital licensing framework should hold companies accountable for governance failures while also empowering licensed practitioners to document and escalate risk.
The third objection is that existing laws are enough. Consumer protection, privacy, product liability, child-safety, labor, and civil-rights laws already matter. But they are often reactive. They intervene after harm, after discovery, after litigation, and after years of public pressure. Professional licensing would add a preventive layer by creating duties before harm becomes a lawsuit.
The fourth objection is that licensing could become exclusionary gatekeeping. This concern is especially important in design, where many excellent practitioners come from nontraditional backgrounds. A licensing pathway should therefore include multiple routes: education, apprenticeship, portfolio review, supervised experience, examination, and continuing education.[16] The goal should be competence and accountability, not prestige protection.
Why Washington State Should Lead
Washington state has the right combination of industry, academic expertise, legal infrastructure, and political urgency to begin this work. Microsoft and Amazon are headquartered here. Adobe, Google, and other major technology companies maintain significant a significant presence. The University of Washington has deep expertise in computer science, human-computer interaction, design, law, and technology policy. The state also has a public interest in regulating the industries that shape its economy and its residents’ lives.
The first step should be a policy brief that establishes the evidentiary case, defines the scope of covered products and roles, proposes a licensing body, and outlines competency standards, insurance requirements, whistleblower protections, and enforcement mechanisms.[17] That brief should be developed with academics, product practitioners, child-safety experts, consumer-protection lawyers, accessibility experts, and legislators.
The second step should be engagement with the Washington State Attorney General’s Office, especially consumer protection and technology policy staff, to identify evidentiary gaps and enforcement needs. The third should be coalition-building with practitioners who understand product development from the inside and with advocates who understand the harms from the outside. The fourth should be a legislative strategy aimed at a realistic bill timeline, beginning with study, hearings, and a narrowly scoped pilot framework for high-risk digital products.
Conclusion: We Have Done This Before
We are at a juncture. Social media harms have already shown how digital products can affect children, families, schools, public health, and civic trust.[18] Artificial intelligence will intensify the same problem by making digital systems more powerful, more opaque, more persuasive, and more difficult for ordinary users to understand or contest.[19] People feel scared, helpless, and angry because the pace of deployment has outstripped public accountability.[20]
But we have been here before. A century and a half ago, rapid technological change in the built environment produced new risks that the public could not manage alone. Buildings collapsed. Fires spread. Workers and residents were injured and killed.[21] In response, society created building codes, professional licensing, inspection regimes, standards of care, liability insurance, and safety enforcement. The system is imperfect, but it saves lives.[22]
Digital infrastructure now deserves the same serious consideration. Product designers and other digital professionals should not be left to rely on private conscience inside companies whose incentives may conflict with public safety. They need enforceable duties, independent standing, and protection when they raise alarms. The public needs standards before harm becomes inevitable.
We can do this again. We must do it again. The place is Washington state. The time is now.
_____________________________
[1] The Standard of Care: How It Is Applied, American Institute of Architects.
[2] The History of Occupational Licensing in the US, Hoover Institution; The History of Occupational Licensing, National Bureau of Economic Research, Working Paper 33580.
[3] Tacoma Narrows Bridge History: The 1940 Bridge, Washington State Department of Transportation; Vertical and Torsional Vibrations Before the Collapse of the Tacoma Narrows Bridge in 1940, Journal of Fluid Mechanics, Cambridge University Press; Tacoma Bridge Collapse: The Wobbliest Bridge in the World? (1940), British Pathé newsreel footage; The Tacoma Narrows Bridge Disaster: A Short Documentary, Fascinating Horror.
[4] Tacoma Narrows Bridge History: Other Bridges' Failures, Washington State Department of Transportation.
[5] Ten Engineering Disasters of the Last 100 Years, American Society of Mechanical Engineers; Disastrous Engineering Failures Due to Ethics, Case Western Reserve University.
[6] Psychological Perspectives on AI Safety and Integration, American Psychological Association.
[7] How Matthew Bergman Is Taking On Social Media Giants, TIME.
[8] People of the State of California v. Meta Platforms, Inc., Less-Redacted Complaint, Office of the California Attorney General.
[9] The Social Media Trials and Section 230, Platformer.
[10] Best States: New Mexico Rankings, U.S. News & World Report.
[11] Statement on Today's Landmark Verdicts, Overturned (Substack).
[12] What Is Stop Work Authority? Airswift.
[13] What Is the Standard of Care for an Architect? Rider Levett Bucknall; The Three Cs Every Architect Should Rely Upon to Minimize the Risk of Liability, L&B Law.
[14] Should UX Researchers Need a License? UX Collective; About the Journal: International Journal of Design Creativity and Innovation, Taylor & Francis Online.
[15] On the empirical literature on the costs and benefits of occupational licensing, see note 2 above.
[16] ARE vs. PE Professional Licensure (forum discussion), Archinect.
[17] Washington's existing professional-engineer licensure requirements offer one procedural model. See Washington State PE License Requirements, Study for FE.
[18] What Will It Take to Get AI Out of Schools? The New Yorker (via Substack).
[19] The Evolution and Impact of AI Integration in Modern Workplaces, PubMed Central.
[20] Why We Fear Diverse Intelligence Like AI, Noema Magazine; Timely Topic: An Expert's Take on Why We Should Not Fear AI, UT Dallas Magazine, April 7, 2025; Diffusion of Innovations, Wikipedia.
[21] Building Collapse Liability: Who Is Responsible? Santoro & Gray.
[22] N.Y.C. Structural Safety Investigation Over Office-to-Residential Conversions, The New York Times, July 10, 2026.
