UNAUTOMATED — THE HUMAN INTELLIGENCE PROJECT

Automate the work. Never automate the human.

HUMAN JUDGMENT. REAL CONSEQUENCE. A BETTER FUTURE.

  • DISCERNMENT. Examine what is true.
  • PRUDENCE. Decide what is worth doing.
  • WISDOM. Act with good judgment.

UNAUTOMATED.ORG

The Unautomated Brief

A daily place to think clearly about AI and the people living with it.

Independent, source-backed commentary on developments that matter—read through the lens of human judgment, accountability, and care.

Published commentary

Latest Briefs

We publish when there is a significant, verifiable development worth discussing—not simply to fill a calendar.

Four people at a table discussing layered blue and violet geometric forms projected above them.
Sep 26, 20267 min read

The AI Debate Needs Responsible Questions, Not a Single Camp

NPR’s guide to the competing factions in the AI safety debate is a reminder that public disagreement is real. It is not a reason to stop asking who bears responsibility for the systems being built and used.

Read the Brief
Four people at a table discussing a projected field of blue and violet geometric forms.
Sep 25, 20267 min read

An Agent’s Identity Is Not Accountability

NIST NCCoE’s work on AI-agent identity and authorization is an important security question. It does not remove the need for a person or institution to remain responsible for what an agent is allowed to do.

Read the Brief
A magnifying glass examining a small paper document on a table.
Sep 24, 20268 min read

An AI Action Plan Is Not Yet an Accountable Justice System

The UK Ministry of Justice has added a fourth AI priority for emerging risks. The update is consequential, but an action plan still has to show how accountability will work when justice decisions affect real people.

Read the Brief
A small group of people in discussion around a table beneath layered blue and violet abstract panels.
Sep 22, 20268 min read

A Partnership Is Not Yet a Chain of Responsibility

The UK and United States announced work on an AI and Autonomy partnership. The ambition is consequential; the public announcement is not yet evidence of adopted governance.

Read the Brief
Abstract editorial illustration of a small gold star above a glowing horizon and dark perspective grid.
Sep 21, 20266 min read

A Notification Is Not Yet Shared Responsibility

Reuters reports that the United States proposed an AI incident-notification mechanism in talks with China. Openness matters, but a proposal is not yet shared responsibility.

Read the Brief
Editorial image of colleagues examining a decision together around a table with a laptop, illustrating the human work required to make responsibility real.
Sep 15, 20267 min read

A Call for AI Rules Is Not Yet a System of Responsibility.

Associated Press reporting shows technology leaders urging stronger AI oversight as federal action remains slow. A public call for rules matters, but responsibility still has to work in practice.

Read the Brief
Colleagues in discussion around a table with a laptop.
Sep 14, 20267 min read

Human Oversight Is Not a Personality Trait

AP reporting renewed the debate over AI guardrails. The practical question for every organization is not whether its people mean well, but whether its systems make responsible human judgment possible.

Read the Brief
Colleagues in discussion around a table with a laptop.
Sep 13, 20267 min read

Slowing Capability Gains Is Not a Pause. It Is a Test of Responsibility.

Reuters reported that Anthropic CEO Dario Amodei called for a slower pace of frontier-AI capability gains. The leadership question is whether safeguards and human judgment are keeping up.

Read the Brief
A magnifying glass examining a small paper document on a table.
Sep 12, 20268 min read

An AI Audit Is Not a Stamp of Approval

California’s new framework for independent AI review is a useful reminder: an audit can strengthen accountability, but it cannot transfer a leader’s responsibility to a report.

Read the Brief
A conceptual lever marked AI against a light background, representing choices about how artificial intelligence is used.
Sep 11, 20266 min read

Responsible AI Still Needs a Human Owner

Microsoft’s latest transparency report points to a practical truth: responsible AI is not a promise a system can make on its own. It is a practice people must lead.

Read the Brief
    The Unautomated Brief | UNAUTOMATED