# Humanity Protocol Founder Says Workplace AI Needs Human Accountability

By Sonny Jang

Canonical URL: https://www.tokenpost.com/news/people/33606
Published: 2026-10-17T00:39:27.000Z
Updated: 2026-10-17T00:39:27.000Z
Section: People

> Terence Kwok says agents can assign work and flag missed deadlines, while people should remain responsible for pay and promotion decisions.

Companies adopting workplace AI agents should keep people accountable for pay and promotion decisions, even as software takes on more workplace tasks, Humanity Protocol founder Terence Kwok says.

Kwok says agents can coordinate work by assigning tasks or flagging missed deadlines. He argues that decisions about employees’ pay and promotions should remain with a named person, even when software supports the process.

“I still want a named person who can stand behind the final result,” Kwok said. In his view, that person must answer for the outcome rather than leave responsibility with an agent that helps coordinate work.

The same boundary shapes Kwok’s view of job mobility. He envisions a worker’s personal agent carrying preferences and professional history from one job to another, while company records and access rights remain with the employer. That distinction separates information tied to a person from files and permissions created or controlled by the company.

Technical identity controls can show how an agent uses workplace systems. Microsoft Entra gives agent identities separate accounts, permissions and sponsors, and sign-in and audit logs can record their activity. The records can identify which agent accessed a system and what it did, but they do not determine who should be accountable for a decision about an employee.

In Kwok’s framework, access controls and human responsibility answer different questions: what the software may do, and who stands behind its results. An agent can have a defined identity and permissions while a person remains responsible for deciding whether its work should affect an employee.

When an agent is retired, Kwok recommends revoking its credentials, tokens and permissions, closing its connections to company systems and preserving a record of its actions. Those steps address the agent’s access and history; they do not replace a named human decision-maker.

Humanity Protocol was introduced in 2023 as a Web3 identity system using palm recognition to verify people in applications. Kwok’s remarks focus on workplace governance and do not announce a Humanity Protocol product for managing workplace AI agents.