On Mon, Oct 27, 2025 at 1:30 PM Jeremy Stanley <[email protected]> wrote:
>
> On 2025-10-27 09:34:03 -0700 (-0700), Alan Coopersmith wrote:
> [...]
> >> The vendor was contacted early about this disclosure but did not
> >> respond in any way.
> [...]
>
> With the flood of dubious reports being submitted by anyone who can
> thumb some words into an LLM prompt and not bother to check the
> results for hallucinated nonsense, I've taken to ignoring or
> summarily closing such submissions to projects I work on as not
> worth my time to respond. This is probably yet another sign that the
> CVE system needs an overhaul or it's going to get ignored when it
> becomes as overwhelmed with "AI noise" as everything else (not
> saying these reports were necessarily machine-generated, but it's
> reaching the point where open source projects with limited resources
> have no choice but to silently bin such nonsense to /dev/null).

cURL is fed up with the LLM nonsense, too.  cURL requires the source
of a vulnerability report be stated because the project was being
overrun with false positives and low quality bug reports from AI
generated slop.  See "AI guidelines" (May 2025),
<https://curl.se/mail/lib-2025-05/0013.html> and
<https://github.com/curl/curl/pull/17325>.

And the IETF is also concerned about submissions curated from LLMs.
See "BCP 78 policy / copyright / Generative AI / LLM .. is there a
FAQ?" (August 2025),
<https://mailarchive.ietf.org/arch/msg/ietf/ZAwDLUWAQ-iU2u6vVpw5IeW7g-E/>.

Jeff

Reply via email to