Daniel P. Berrangé <[email protected]> writes:

> On Tue, Jun 16, 2026 at 07:06:23PM +0200, Christian Borntraeger wrote:
>> Am 29.05.26 um 11:46 schrieb Paolo Bonzini:
>> > -Exceptions do not remove the need for authors to comply with all other
>> > -requirements for contribution.  In particular, the "Signed-off-by"
>> > -label in a patch submission is a statement that the author takes
>> > -responsibility for the entire contents of the patch, including any parts
>> > -that were generated or assisted by AI tools or other tools.
>> > +.. code-block:: none
>> > +
>> > +     AI-used-for: tests, docs
>> > +     AI-used-for: code
>> > +     AI-used-for: code (refactoring)
>> > +     AI-used-for: code (prototype)
>> > +     AI-used-for: research
>> > +
>> > +``AI-used-for`` should not be included for "background" usage such as
>> > +autocomplete or obtaining a pre-review of the patch.
>> 
>> So what about using AI for security scanning? So how do we want to treat
>> a patch from a human that is based on an AI report.
>> And if ok, would we then add something like
>> 
>> Reported-by: Claude, chatgpt whatever?
>
> I see no need to provide advertizing for these tools.
>
> Our security disclosure rules require that the submitter acknowledge
> they have reviewed any LLM output themselves before reporting.  So
> from that POV it is the human who gave us the report and whom
> deserves the credit, not any tool or vendor.
>
> FWIW, although pretty much every security report recently smells
> strongly of LLM (highly structured markdown with headings that
> look the same from all reporters), almost none of them credit the
> specific tool used anyway.

We have created a new label Audit Tooling: AI in the issues but the most
interesting thing about that is to see how AI does compared to existing
fuzzing and static analysis techniques. The underlying model isn't super
interesting here.

>
> With regards,
> Daniel

-- 
Alex Bennée
Virtualisation Tech Lead @ Linaro

Reply via email to