Human-in-the-loop laws are arriving. They are really about ownership.

On 30 September 2026, California Governor Gavin Newsom signed SB 947, the so-called "No Robo Bosses Act", according to coverage including HR Dive. As reported, the law bars employers from relying solely on an automated decision system to discipline or fire an employee, requires a human to review and corroborate such decisions, and requires written notice to affected workers. It is reported to take effect on 1 July 2027.

The details matter for HR teams in California. The principle matters for everyone who uses software to help make decisions.

The principle: decisions need a human owner

Laws like SB 947 encode something good teams already practise: software can inform a decision, but a named person must own it. When "the system flagged it" is the only explanation for an outcome, nobody is accountable, and nobody can explain or correct it.

This is the same idea we apply to work generally. An action owned by "the team" tends not to happen; see social loafing. A decision owned by "the algorithm" has the same problem, with higher stakes.

What this looks like in practice

For any decision where software plays a part, whether a hiring screen, a performance flag or an automated customer refund, ask:

Who owns this decision? One named person.
What evidence did they consider beyond the system's output?
Where is that recorded?
Who can the affected person talk to?

Those questions turn an automated output into an owned decision.

Boards and decisions

A board does not make decisions, and it should not. But it can make the ownership of decisions visible. When a decision is pending, it sits as an action in someone's court. When it is waiting on information, it sits in Other with the person who owes it. When it is done, the record shows who decided. See ownership vs responsibility.

Beyond California

Other jurisdictions are debating similar rules for automated decisions at work, and the EU's AI Act already treats some workplace AI uses as high risk. Whatever the legal timeline where you work, building the habit of human ownership now is good practice and good preparation. For how agents are changing day-to-day work, see always-on AI agents and ownership.