I don't think "usage" is exactly the metric they're going for, more like "usage in line with our developmental strategy." Transcripts of people using Claude to write code are probably far more valuable to them than transcripts of OpenClaw trying to set up a calendar invite.

I mean, they don’t train on your data unless you have the setting enabled. Do you really think they are reading your prompts at all? Free inference providers sure, but Anthropic?