Published Works
Tech Shall Not Kill
Building the Consequence Layer for AI Agents
By Edward Fassio
Between a proposed action and its real-world consequences, there must be more than a rule.
Build the consequence layer
Tech Shall Not Kill explores independent controls between a proposed agent action and real-world execution. The fictional Harborline Services illustrates how permissions, budgets, approvals, emergency stops, durable records and human sponsorship make responsibility visible.
From boundaries to action
Permissions and approvals
Separate what an agent proposes from what it is permitted to execute. Independent controls and human sponsorship give boundaries practical meaning.
Budgets and emergency stops
Put limits around action and retain a way to stop it. The fictional Harborline Services makes these operational questions concrete.
Durable records
Keep a record of decisions and actions so responsibility remains visible. Plain-language field notes connect the controls to everyday judgment; no coding is required.
Who this is for
For readers responsible for the consequences of AI-enabled work: leaders, sponsors and teams who need to understand how boundaries become operational controls. Plain-language field notes; no coding required.
The series connection
The fourth book in the Agentic Frontier Series. It extends the conversation from human accountability and organizational collaboration to independent controls between proposed actions and execution.
About the author
Edward Fassio writes about AI collaboration, delegated authority and human accountability.
Get it on Amazon