Licensing Computer Software & Hardware Engineers Who Create AI Applications: A Necessity in a Transforming World

Imagine a world where an untested AI system determines who gets a mortgage, which job applications get seen, or how law enforcement is deployed. The consequences of poorly designed or biased algorithms have already led to wrongful arrests, discriminatory hiring practices, and fatal accidents in autonomous vehicles. Without licensing, anyone with the right skills—but not necessarily the proper ethical training—can create AI models that impact millions.

The dangers aren’t just theoretical. Real-world failures—like biased facial recognition wrongly identifying innocent people or self-driving cars misjudging pedestrian safety—underscore the urgent need for accountability. Engineers are held responsible when bridges collapse, and doctors face scrutiny when medical treatments go wrong. But when AI harms people, who is accountable?

Accountability is crucial in an era where artificial intelligence (AI) applications significantly influence human lives. Licensing computer software engineers who develop AI applications for commercial and public use is no longer a mere consideration but a pressing necessity.

As industries mature, so do the laws, regulations, and expertise of those who practice and enforce them in the field. The precedent exists for licensing in most engineering disciplines, as well as architects, lawyers, doctors, dentists, and hair stylists. All these professions involve safeguarding human lives, persons, or property. However, with each disruptive event, e.g., the personal computer, the World Wide Web, and now AI, society realizes that law always lags behind technology. Nevertheless, it's time that computer engineers join the ranks of other engineering professions and formally recognize the value of apprenticing, mentoring, and licensing for the common good of all.

Below are five compelling arguments for such licensing, each supported by examples and counterarguments.

1. Protecting Public Safety

AI systems are increasingly deployed in safety-critical domains such as autonomous driving, healthcare, and infrastructure management. Comparatively, almost everyone in healthcare is licensed or certified except the non-clinical technology staff, leaving a gaping hole for the commission of errors or exploitation of security protocols. For instance, a malfunctioning AI algorithm in an autonomous vehicle could result in catastrophic accidents. Even without the nefarious use of AI tools, USA-based health record systems are breached, sometimes ransomed, daily due to lax security measures despite available remediation tools.

Licensing software engineers ensure that individuals and their employers developing these applications possess the necessary expertise and adhere to rigorous ethical and technical standards.

Example pro:?The aviation industry requires stringent licensing for pilots and maintenance engineers to safeguard lives. Similarly, licensed AI engineers could mitigate risks in high-stakes environments like healthcare, smart cities, and autonomous vehicles.

Example con: Critics argue that licensing might stifle innovation and deter talented individuals who prefer less regulated environments. However, safety should supersede unbridled innovation, especially when lives are at stake.

2. Establishing Accountability and Ethical Standards

AI systems often perpetuate biases, sometimes unintentionally. Licensing engineers would mandate ethical training and accountability, ensuring developers are equipped to mitigate biases in AI design. For example, facial recognition technologies have demonstrated racial and gender biases, leading to wrongful arrests and discrimination, and several pilot HR applications were found to be biased.

Example pro: Licensed engineers could be held legally accountable for deploying biased systems, encouraging them to design more equitable solutions.

Example con: Opponents might argue that accountability should lie with organizations rather than individual engineers. However, shared accountability between organizations and licensed engineers creates a dual safeguard against unethical practices.

3. Ensuring Compliance with Regulatory Standards

Governments worldwide are beginning to regulate AI applications, but enforcement becomes challenging without licensed professionals. Licensed engineers could act as gatekeepers, ensuring compliance with data protection laws, such as the General Data Protection Regulation (GDPR), and sector-specific AI guidelines.

Example pro: A licensed engineer developing a medical diagnostic AI must comply with healthcare regulations, reducing malpractice risks.

Example con: Critics might contend that regulations alone suffice. However, rules without competent professionals to enforce them can lead to superficial compliance rather than meaningful adherence.

4. Professionalizing the Field of Computer Engineering and AI Development

Licensing elevates professions like law, medicine, or architecture. It instills a sense of responsibility and pride among practitioners, potentially attracting more talent to the field.

Example pro: The medical field’s licensing requirements ensure high standards, earning public trust. Licensed AI engineers could similarly bolster public confidence in AI systems.

Example con:?Some might argue that AI's rapidly evolving nature makes standardization impractical. However, licensing bodies can adopt adaptive models to keep pace with technological advancements, which are just as dramatic in already licensed fields involving new power grids and commercial construction designs.

5. Minimizing the Risk of Malicious AI Applications

Licensing can deter malicious actors who exploit AI for harmful purposes, such as deepfake propaganda or cyberattacks. Licensed engineers would be required to adhere to ethical codes, reducing the likelihood of intentional harm.

Example pro: A licensed engineer would face severe repercussions for knowingly developing AI systems used for fraudulent activities.

Example con: Detractors might argue that malicious actors would circumvent licensing systems. While no system is foolproof, licensing raises the barrier to such activities and aids in tracing accountability.

Conclusion

Licensing software engineers who develop AI applications is essential to ensure safety, accountability, compliance, professionalism, and ethical responsibility. While concerns about stifling innovation and implementation challenges exist, these are outweighed by the potential for catastrophic consequences in unregulated AI development. Policymakers, industry leaders, and academic institutions must collaborate to create a licensing framework that evolves with technology while safeguarding the public interest.

References

Binns, R. (2018). Fairness in machine learning: Lessons from political philosophy. Proceedings of the 2018 Conference on Fairness, Accountability, and Transparency.

Goodman, B., & Flaxman, S. (2017). European Union regulations on algorithmic decision-making and a "right to explanation." AI Magazine, 38(3), 50-57.

O'Neill, C. (2016). Weapons of math destruction: How big data increases inequality and threatens democracy. Crown Publishing Group.

Russell, S., & Norvig, P. (2021). Artificial intelligence: A modern approach (4th ed.). Pearson.