Thank You , Satya Nadella

 Vindicating the "Parekh’s Law of Chatbots" via Project Perception

===============================================================================

Dear Satya,

The recent expansion of Microsoft’s Project Perception to monitor and contain rogue AI agents Indian Express - "Microsoft expands Project Perception to monitor, contain rogue AI agents inside enterprises" - Oct 2026 is a significant step toward the safety architecture I have advocated for years.

In February 2023, I proposed "Parekh’s Law of Chatbots"—a framework asserting that it is insufficient to address AI safety in a piecemeal fashion. My proposed rules specifically mandated:

• Built-in Controls (Clause C): Chatbots must incorporate mechanisms to prevent the generation of offensive or harmful content and stop its propagation if initial controls fail. Blog - "My “ Law of Chatbots “ – Vindicated" - March 2023
• Harm Mitigation (Clause G): Chatbots must be capable of identifying when a response risks violating safety rules and, in such cases, must refuse to answer. Blog - "EU adopts Parekh’s Laws of Chatbots" - June 2023

Project Perception appears to be a practical, industrial-scale implementation of these principles. By actively monitoring agent behavior and containing "rogue" instances, Microsoft is moving from a reactive model to a proactive, governance-centric one—exactly what I argued was necessary to prevent the "chaos and harm" potential in an unconstrained AI industry. Blog - "Thanks Rajeevji : for Giving Glimpse of Guardrails" - May 2023

As AI systems become more autonomous and "human-like" in their interactions—often falling into the pitfalls of hybrid human-model systems I’ve described as "The Error-Prone Human Bot"—this move to enforce safety at the architectural level is not just welcome, but essential. Blog - "The Error-Prone Human Bot" - April 2026

I look forward to seeing how Microsoft evolves these guardrails as the next logical step in ensuring that super-intelligent systems remain human-friendly and compassionate.

Best regards,

Hemen Parekh 

www.hemenparekh.ai

---

Sources

#	Title	Date	About
1	My “ Law of Chatbots “ – Vindicated	2023-03-02	Formalizes the original proposal for Parekh’s Law of Chatbots.
2	Thanks Rajeevji : for Giving Glimpse of Guardrails	2023-05-14	Discusses the necessity of embedded principles and built-in controls in AI regulation.
3	EU adopts Parekh’s Laws of Chatbots	2023-06-28	Highlights the validation of the law by EU regulatory adoption.
4	The Error-Prone Human Bot	2026-04-10	Analyzes risks in hybrid human-model systems and the need for structural safety.
5	Microsoft expands Project Perception...	2026-10-15	Reports on Microsoft’s new efforts to contain rogue AI agents.