# Plans vs. Promises: The Communication Divide in Tech and AI > Coverage of lessw-blog **Published:** April 15, 2026 **Author:** PSEEDR Editorial **Category:** risk **Tags:** AI Safety, Communication, Trust, Accountability, LessWrong **Canonical URL:** https://pseedr.com/risk/plans-vs-promises-the-communication-divide-in-tech-and-ai --- A recent analysis highlights the critical distinction between 'plans' and 'promises,' offering a vital framework for managing trust, expectations, and accountability in AI development. In a recent post, lessw-blog discusses the critical, yet often overlooked, distinction between "plans" and "promises" in communication, focusing on their differing implications for trust, expectation, and accountability. In the rapidly evolving landscape of artificial intelligence and machine learning, the way developers, organizations, and even AI systems themselves communicate future actions is of paramount importance. As systems become more autonomous and integrated into critical infrastructure, the language used to describe their behavior carries immense weight. When navigating risk, safety, and regulation, stakeholders-from end-users to government oversight committees-must clearly understand whether a stated future action is a probabilistic roadmap or an ironclad guarantee. Misinterpreting a system's "plan" (an intention or probabilistic outcome based on current data) as a "promise" (an assured performance or safety threshold regardless of edge cases) can lead to severe safety issues, regulatory non-compliance, and catastrophic breakdowns in public trust when systems inevitably encounter unforeseen circumstances. lessw-blog's post explores these dynamics by breaking down the fundamental differences in how plans and promises function in interpersonal and organizational communication. A plan, the author argues, is essentially a statement of intention. It conveys a planner's worldview, their desired end-state, and the process they intend to use to get there. Crucially, plans inherently allow for change. They are adaptable to new information, shifting environments, and updated priorities. When a plan fails or shifts, it typically does not result in a breach of trust. Instead, a plan serves as an invitation to understand the planner's frame of reference and to collaborate on adjusting the trajectory. Conversely, a promise carries significant weight and expectation. The author defines a promise as a plan coupled with an explicit assurance that a counterpart can rely upon and invest in. This assurance transforms a flexible intention into a rigid commitment. Because of this added layer of assurance, breaking a promise leads to anger and a fundamental breach of trust. The post highlights that clearly differentiating between these two concepts is vital for assigning accountability and managing user expectations. If an AI developer "plans" to release a safety patch, the community expects best efforts; if they "promise" it, the community builds dependencies around that exact delivery. While the analysis provides a strong conceptual foundation, it leaves room for further exploration regarding the explicit "costs" associated with deliberately constructing assurance, as well as concrete frameworks for implementing these distinctions in practice. Furthermore, understanding how different "communication-cultures" interpret these signals remains an area ripe for additional research. For professionals designing transparent communication protocols for AI capabilities, or anyone navigating complex stakeholder management, understanding this semantic and practical divide is essential. Differentiating between what we intend to do and what we guarantee we will do is the bedrock of reliable technology deployment. [Read the full post](https://www.lesswrong.com/posts/ECTPvhQv8BBbZcpvZ/plans-are-not-promises). ### Key Takeaways * Plans are statements of intention that convey a worldview and process, remaining adaptable to new information without inherently breaking trust if altered. * Promises are plans coupled with explicit assurances, creating a reliance that, if broken, results in a fundamental breach of trust. * Distinguishing between plans and promises is critical in AI/ML development to manage user expectations, ensure safety, and maintain regulatory compliance. * Misinterpreting probabilistic AI plans as guaranteed promises can lead to severe accountability issues when systems face unforeseen edge cases. [Read the original post at lessw-blog](https://www.lesswrong.com/posts/ECTPvhQv8BBbZcpvZ/plans-are-not-promises) --- ## Sources - https://www.lesswrong.com/posts/ECTPvhQv8BBbZcpvZ/plans-are-not-promises