RANT Roundtable in partnership with WTW

It’s easy to see why so many businesses are deploying generative AI across their systems.There’s a ton of drudge work keeping the wheels of commerce turning which large language models are tailor-made to carry out. Automating unpopular and time-consuming tasks doesn’t just unleash your staff to be more productive – it probably makes them happier, and that’s going to make them more loyal.

For those businesses reliant on Microsoft technologies, the pressures to adopt AI aren’t just internal. The trillion-dollar behemoth is adamant that its Copilot tool will help your business answer all its major questions, and is increasingly keen for its clients to roll it out across their networks.

The only snag: Copilot – or any LLM let loose on a corporate system – will ingest everything it can, storing every bit and byte of data, ready to spit them back out at unpredictable times and in contexts you can’t even imagine. Microsoft says you can prevent this by tagging files you don’t want Copilot to be able to access, but who’s got time to go through every digital file on every server and endpoint across the entire enterprise, work out which are OK for Copilot to read and which aren’t, and give each the appropriate flag?

Preventing AI from accessing sensitive documents is going to become a key priority for those security teams who aren’t already losing sleep over the risks this raises. Reining in the AI beast will be difficult, but it need not be impossible. Join a select group of your fellow senior security leaders, and representatives from our hosts, WTW, to discuss how you can quickly, easily and relatively painlessly start to fence in your wandering AI.