Tom Loosemore

It’s one of those fascinating periods when different worlds collide.

I had two conversations this last week about the need for lighter-weight legal consents for acting on behalf of someone else. One conversation was with US-based Dave Guarino about the need for standards around the use of AI agents that help people access public services. Agents are going to be a really big deal for accessing public services – not least joining them up – but how does a service decides if an AI agent really is acting legitimately on my behalf?

The other conversation was with folks from the Connection Project about how Lasting Power of Attorney is simply too all-or-nothing to adapt to the evolving needs of people who need support from others to access essential digital services like banking, telecoms or utilities.

Both scenarios need faster, lighter, more flexible ways to prove that that you (you being a human or an AI agent!) has been given proper consent to act on someone else’s behalf.

A truly AI-friendly innovative government should be experimenting with digitally-native legal consents which could vary both in duration and scope.

We need legislative sandboxes, not just ones for regulation or technology. Pick a willing village…