# Building Technology to Drive AI Governance: A Third Path for Engineers > Coverage of lessw-blog **Published:** February 18, 2026 **Author:** PSEEDR Editorial **Category:** risk **Content tier:** free **Accessible for free:** true **Word count:** 485 **Tags:** AI Governance, AI Safety, Tech Policy, Engineering Strategy, Regulation **Canonical URL:** https://pseedr.com/risk/building-technology-to-drive-ai-governance-a-third-path-for-engineers --- A new framework proposes that engineers can best influence AI safety not by lobbying or internal compliance, but by building external tools that make governance technically feasible. In a recent post, **lessw-blog** discusses a strategic framework for technically skilled individuals who want to influence AI governance without abandoning their engineering roots. The analysis addresses a common dilemma facing the technical community: the perceived binary choice between working on internal safety teams at major AI labs or pivoting entirely to policy and advocacy work. **The Context: The Limits of Current Approaches** For engineers concerned with the trajectory of Artificial Intelligence, the standard career paths often feel limiting. Working internally on alignment or safeguards allows for direct technical contribution, but these efforts are frequently constrained by leadership priorities, commercial incentives, and the competitive race to release models. Conversely, moving into direct policy involvement often requires engineers to sacrifice their comparative advantage-their technical skill set-to operate in a crowded, slow-moving political domain where they may lack leverage. **The Gist: Technology as a Governance Lever** The post argues for a "third path": building technology that drives governance by fundamentally shifting the environment in which AI development occurs. Rather than relying solely on voluntary internal compliance or top-down regulation, this approach focuses on creating tools that alter information availability, economic incentives, and strategic options. The author illustrates this concept using a historical analogy regarding methane leak regulation. For years, regulating methane emissions was difficult because leaks were invisible and hard to quantify. The development of infrared imaging technology changed the landscape; it made leaks visible and measurement cheap. This technological advancement did not just support regulation-it made regulation feasible where it previously was not. The post suggests that the AI sector requires similar "infrared cameras"-technologies that create visibility and accountability. **Mechanisms for Change** The analysis highlights specific mechanisms where technical builds can force governance outcomes: * **Measurement:** Creating independent tools to verify model capabilities, compute usage, or data provenance. If external parties can measure risk or non-compliance, it forces accountability that internal teams cannot easily ignore. * **Cost Reduction:** (While the summary notes this section was cut off, the implication is clear based on the context). Developing tools that make safety evaluations or compliance checks cheaper and faster removes the economic friction of being safe, altering the cost-benefit analysis for developers. This perspective is particularly relevant for the PSEEDR audience, as it reframes "DevTools"-often seen merely as productivity boosters-as potential instruments of policy enforcement. By building the infrastructure of measurement and verification, engineers can create the conditions necessary for effective governance to take root. For a deeper understanding of how technical builds can shift the regulatory landscape, we recommend reading the full analysis. [Read the full post at LessWrong](https://www.lesswrong.com/posts/weuvYyLYrFi9tArmF/building-technology-to-drive-ai-governance) ### Key Takeaways * Technically skilled individuals often face a false dichotomy between internal safety work (constrained by profit) and policy work (lacking technical leverage). * A 'third path' exists: building external technologies that shift the information and incentive landscape of AI development. * Historical examples, such as infrared imaging for methane leaks, demonstrate how technology often precedes and enables effective regulation. * Key mechanisms for this approach include 'Measurement' (creating visibility) and lowering the costs of safety compliance. * This approach allows engineers to utilize their comparative advantage to make governance feasible rather than just advocating for it. [Read the original post at lessw-blog](https://www.lesswrong.com/posts/weuvYyLYrFi9tArmF/building-technology-to-drive-ai-governance) --- ## Sources - https://www.lesswrong.com/posts/weuvYyLYrFi9tArmF/building-technology-to-drive-ai-governance