Stop Writing AI Policies. Build This Instead.

As we start to see AI being rolled out everywhere and everybody experimenting with it, we're starting to sort of see different strategies come to light in different organizations and, and different reactions to risk.

And I've been really interested in just observing some of those different approaches.

And what I would say is I've seen a lot of those sort of hundred-page policy documents, um, and I've seen a lot of steering committees established, but essentially what I haven't seen is the risks being addressed.

So it kind of feels like there's been a lot of tick box exercises, but without actually making sure that these policies are actually adopted.

So my message today is to stop writing policy after policy after policy and start building real protection.

Now, the reality is, in most organizations, teams are out there experimenting.

We have so many available tools at our fingertips, and most of those are really cheap or sometimes free.

And so you've got employees out there experimenting with the different tools, which, you know, I think is a great thing, but there does need to be some management around the risk.

There does need to be some governance and protection.

Now, in many cases, I would say that a lot of those employees aren't even telling their employers that they're using them.

So you kind of have a big blind spot, um, and that really does expose some risk.

You know, you can imagine probably at least half or a third or maybe two-thirds of your employees potentially are putting company information into AI tools that you don't even know about, and you have no idea what the protection is around that data.

These tools are not vetted.

These tools are not company approved.

And what is our instinctive response when we see this happen? What we do is we go and put another hundred-page policy in.

We go and send another company-wide communication in.

We start to restrict as much as possible, making it so restrictive sometimes that actually you're hindering progress and actually falling behind.