About the Authors

  • The Authors and Contributors of "Patent Docs" are patent attorneys and agents, many of whom hold doctorates in a diverse array of disciplines.
2018 Juristant Badge - MBHB_165
Juristat #4 Overall Rank

E-mail Newsletter

  • Enter your e-mail address below to receive the "Patent Docs" e-mail newsletter.

Contact the Docs


  • "Patent Docs" does not contain any legal advice whatsoever. This weblog is for informational purposes only, and its publication does not create an attorney-client relationship. In addition, nothing on "Patent Docs" constitutes a solicitation for business. This weblog is intended primarily for other attorneys. Moreover, "Patent Docs" is the personal weblog of the Authors; it is not edited by the Authors' employers or clients and, as such, no part of this weblog may be so attributed. All posts on "Patent Docs" should be double-checked for their accuracy and current applicability.
Juristat #8 Overall Rank


« Sequoia Technology LLC v. Dell Inc. (Fed. Cir. 2023) | Main | Stanford Asks Supreme Court to Revisit Subject Matter Eligibility on Diagnostic Claims »

April 30, 2023


I would point out that any legislation on the EU side as to "harmful uses" may well NOT be amenable to US legislation, given our First Amendment.

The entire notion of "spreading disinformation" SHOULD BE SEEN as a logistical/political/ideological swampland.

As we 'learn' more and more about such things as COVID and the like, what was largely denigrated as "conspiracy theory" and "disinformation" has turned out to be plainly factual information.


Isn't "disinformation" meaningless for generative AI ? For a generative AI, truth does not exist.

As to "harmful uses", I assume a generative AI has built-in moderation algorithms which block offensive language and the like for obvious reasons.


I cannot agree.

Generative AI may well yield a result that is a factual dissertation, and thus "truth" very much is reachable.

As for "built-in," these have been labeled "guard rails," and -- like it or not -- have been shown to not only be rather easily circumventable, but with the proliferation of AI tools (including ones that are 'let loose' and interact outside of guard rails and bootstrap with OTHER AI tools, there is no doubt whatsoever that "harmful uses" will abound.

Quite in fact, several notable pioneers in the AI field have been quite public about this in just the last few days.

There is a problem with your conclusion that truth is reachable with generation AIs : you have no access to the sources, the output is not vérifiable.


Access to source is not a constraint to verifiability of outcome.

Quite in fact, most accepted norms of verifiability encourage second source verifications.

Regardless - this has zero effect on whether or not truth is in fact reachable. There is NO problem with my conclusion.

The comments to this entry are closed.

May 2024

Sun Mon Tue Wed Thu Fri Sat
      1 2 3 4
5 6 7 8 9 10 11
12 13 14 15 16 17 18
19 20 21 22 23 24 25
26 27 28 29 30 31