Co-author of Judgment Capital: What Boards Must Preserve When AI Does the Thinking. Organisational change and IT implementation consultant, ex-Kearney.
A governance team on Reddit described what happened when they were asked for an inventory of every AI tool employees actually use. The approved-vendor list supports procurement and shows nothing about use.
https://t.co/b3zP80lgFh
Anadolu Agency reported that experts told US lawmakers AI models are now reasoning internally in ways humans cannot observe. The instinctive response is to insist that models be made transparent before we trust them.
https://t.co/b3zP80lgFh
In August a group of researchers, Grace Liu, Brian Christian and colleagues, posted a revised paper on a series of randomised controlled trials with 1,222 participants.
https://t.co/b3zP80lgFh
There is one place where the erosion of a skill under AI has actually been measured, and I think every board approving an AI deployment should know the numbers.
https://t.co/b3zP80lgFh
Luiza Jarovsky made the case in her newsletter recently that AI is governable, because liability law is an old mechanism that already works and can hold companies to account.
https://t.co/b3zP80lgFh
A thread on Reddit's AI governance forum asks a clean question this week. AI gives you sensible advice, you follow it, something goes wrong: who is responsible? The replies treat it as a menu of the person, the vendor, both or neither.
https://t.co/b3zP80lgFh
Any argument that a risk belongs on the register should come with the conditions under which it comes off again. Without them it is a mood, and a CRO would be right to put it in the wastebasket.
https://t.co/b3zP80lgFh
Oksana and I spent a good part of chapter 2 of our book on a study that would settle whether AI is eroding how professionals form judgment. It can never be run, and the reason matters for boards.
https://t.co/b3zP80lgFh
On 9 December 2025 George Botic, then Acting Chair of the PCAOB, the body the US created after Enron to inspect audit firms, gave a speech at the AICPA & CIMA conference in Washington. Most of it defended the auditor's role.
https://t.co/b3zP80lgFh
Picture a model that triages five hundred items and passes twenty to a person for review. The twenty get read, argued over, sometimes overturned.
https://t.co/b3zP80lgFh
On 1 November, Oksana Sivokobilska and I publish Judgment Capital: What Boards Must Preserve When AI Does the Thinking.
It is about the judgment an organisation runs on, and what happens to it when the work that used to train it is automated.
https://t.co/b3zP80lgFh