X has provided data (public) that Media Matter committed fraud to game the system.

How did Media Matters commit fraud?

From the complaint, they tested the truth of X's statements that major brands were protected from toxic content by doing the following:

1) They took an account and had it follow a bunch of other accounts that posted a lot of anti-semitic content, so that the anti-semitic content would show up in their account's feed; and
2) Ran that account through a bunch of refresh cycles so they could speedrun through all the ads to see if major brands showed up next to the toxic content; and
3) When that happened, they took screenshots of the results and posted them on the web.

None of that is fraud. None of it is "gaming the system" - it's exactly what you would do if you wanted to test whether Twitter's algorithms actually worked to prevent major brands' advertising from showing up next to toxic content. And MM proved they didn't - if you're online looking at toxic content on Twitter, Twitter's algorithms can put a major brand ad next to that content.

Musk is suffering the completely foreseeable consequences of his decision to implement a "speech not reach" policy on Twitter without being able to deliver the "not reach" aspect. Old Twitter used to remove toxic content that violated their "Hateful Conduct" rules, and used to remove users that regularly posted such content. Musk's Twitter allows both the content and the users to remain on the site. The "not reach" element presumes that Twitter can keep the toxic content and posters isolated away from the normie users of the site and ensure that major brand advertisers are altogether separated from the Hateful Conduct. The events that precipitated the latest advertiser withdrawal - Musk's exposure to and subsequent retweeting of anti-semitic content and MM's stress test of the advertising algorithms - show that Twitter can't deliver the "not reach" element.