Post by Carey L.

Yesterday, I saw an alarming headline on LinkedIn: #Microsoft was harvesting our Word documents to train Copilot (aka, #ClippyLLM)! And they had opted everyone in, by default, as part of some larger ‘Connected Experiences’ offering. Aaaggghhhhh.. Obviously, this raised the Privacat #CatSignal and I immediately began digging in to figure out what the hell was going on. It turns out, Microsoft’s ‘Connected Experiences’ weren’t actually as ominous as they initially seemed to be, and that Microsoft, was not, in fact, harvesting our Word documents to train Clippy, but Microsoft had nonetheless committed a cardinal sin: They hid important details from their users, opted everyone in by default, and then inconsistently explained themselves across dozens of different documents in the most confusing way possible. In short, they stepped on their own dicks. Again. And so, I talk about that, why it’s bad, and how tech companies can do better. cc: Shoshana Rosenberg Ingrid King Philip M. Cole Pascua 🇨🇦 🇵🇭 🏳️‍🌈Tiffany Walker-Roper, AIGA, GRCP Michael J. Silva Lisa LeVasseur Dagmar Monett Julie Saslow Schroeder, J.D., L.L.M.(Actively searching) Debbie Reynolds Erin Weslock Rosalia Anna D'Agostino Jeff Jockisch