I would still claim two differences on the user agent side of things:
First, any browser can report any user agent they want. There are a number of examples of browsers faking or changing user agents to get around sites that try to differentiate based on those strings.
Second, while any individual web site can implement logic based on the user agent, that's the sole choice of the web operator. By restricting Widevine access, Google is blocking rendering of content on other people's domains.
The non-ISP, in-browser analogy I would use would be if Google decided that in order to render an AMP page in your new browser, you first needed to get their permission. They're not just blocking their own content, they're blocking an entire category of technology.
It's also worth mentioning that even under the user agent analogy, if this headline was, "Google uses user agents to block Firefox from accessing Youtube", pretty much every person on HN would call that anti-competitive behavior worthy of regulation.