While this is
technically true, it's not a correct use of DKIM: DKIM is meant to provide authenticity for the receiving mail server (by verifying the message against a key associated with the sending server), not
nonrepudiation for the sending user.
Many email providers historically used short keys (RSA <512, and later 1024) for DKIM, meaning that lots of emails circa 2012 and older appear to be nonrepudiable but in fact are.
Long-term authenticity of emails is very difficult and is arguably an anti-feature; Matt Green correctly observes[1] that Google should be rotating and publishing its DKIM keys regularly to prevent people from falsely concluding that a DKIM signature is "proof" of a message's authenticity.
[1]: https://blog.cryptographyengineering.com/2020/11/16/ok-googl...