Re: We should require AI disclosure

--===============7752431507778712303==
Content-Type: multipart/alternative;
 boundary=----N6UEHOVTLDB10FEFDLE8LQDNZLIB5T
Content-Transfer-Encoding: 7bit

------N6UEHOVTLDB10FEFDLE8LQDNZLIB5T
Content-Type: text/plain;
 charset=utf-8
Content-Transfer-Encoding: quoted-printable

Hi,=20

It is still not very clear to me what the objective is and how a requireme=
nt to disclose the use of AI achieves it=2E I see these angles:=20

1=2E AI may generate output that is of low quality (incomplete, vague, fal=
se, superfluous)=2E

2=2E AI may generate output that includes parts of other works without att=
ribution and/or permission (from a rights perspective)=2E

I do not see why these problems uniquely exist because of the use of AI, a=
nd why this isn't covered by an author's existing responsibility to ensure =
quality and the pre-conditions for assigning ownership rights to the XSF pe=
r its IPR policy (in particular sections 3=2E1 and 3=2E2)=2E

If you cannot formulate the requirements for quality for humans, then how =
does declaring use of AI make things better? And if you can, and a submissi=
on complies, how does it matter that AI was used in the process? How would =
you handle =E2=80=9CAI-tainted=E2=80=9D submissions differently?

The same holds for potential rights issues=2E The author is still responsi=
ble=2E=20

My concern is that all future submissions will just have this disclosure a=
s boilerplate, and a recipient cannot assess how deep the impact of the use=
 of AI is=2E With the accelerating growth we see in both the capabilities a=
nd the use of AI, this becomes increasingly hard=2E We will not have gained=
 anything substantial=2E In that regard it reminds me of the Evil Bit (RFC =
3514) / Malicious Stanzas (XEP-0076)=2E

The statement you linked to seems like common sense and doesn't actually r=
equire disclosure=2E Consider this document again, but conceptually replace=
 =E2=80=9Cthe use of AI=E2=80=9D with the =E2=80=9Cuse of a keyboard=E2=80=
=9D=2E Does having such a statement change the outcome of our standards pro=
cess?

I think that our author guidelines (XEP-0143) are already clear, and sense=
 all the above boils down to this:

    =E2=80=9CAuthors must understand, verify, and take responsibility for =
every contribution they submit=2E=E2=80=9D (-- ChatGPT with my prompting)

While noting that this stance is not universally accepted (e=2Eg=2E see <h=
ttps://jme=2Ebmj=2Ecom/content/51/4/230>), I think it is suitable for our s=
tandards process=2E Feel free to use this phrase in a concrete proposal, li=
ke a PR to XEP-0143=2E

Cheers,=20

ralphm

Disclosure: this message was manually (glide) typed on a OnePlus 12, using=
 Google Board, in Thunderbird for Android=2E Conversing with AI may have in=
fluenced my thought process=2E Except where explicitly noted, no excerpts o=
f other works were included in this message=2E



On 7 July 2026 17:55:54 CEST, Goffi <goffi@goffi=2Eorg> wrote:
>Hi,
>
>This discussion is stalling=2E
>
>Meanwhile, we've got specification(s?) which have been clearly written wi=
th AI,=20
>and I guess it will be more and more often the case=2E
>
>As a reviewer with my council hat, I would really love to at least have=
=20
>disclaimer when AI is used (is it for writing whole sections, to extract =
a=20
>table, to write example, to check spelling/grammar)=2E
>
>Many big projects have a statement on AI use, e=2Eg=2E, CPython:
>https://devguide=2Epython=2Eorg/getting-started/ai-tools/
>
>I think XSF should have one too=2E
>
>Do we need more discussion on standard, or should the board discuss that =
and=20
>ask a team to work on a AI statement?
>
>Thanks
>Goffi
>
>
>Le lundi 11 mai 2026, 09:50:19 heure d=E2=80=99=C3=A9t=C3=A9 d=E2=80=99Eu=
rope centrale Goffi a =C3=A9crit :
>> Hello everybody,
>>=20
>> I would like to bring a discussion on AI policy=2E We can't really igno=
re=20
>> anymore that modern models have become very capable, and I suspect that=
 they=20
>> are used for spec authoring=2E
>>=20
>> This raises, I believe, copyright issues: if someone use AI to redact a=
=20
>whole=20
>> section of a spec, how can we be sure that it's not an existing specs f=
or=20
>some=20
>> other place, possibly under copyright, that is copied or paraphrased? H=
ow=20
>can=20
>> an author guarantee that it's original work (hint: they can't)?
>>=20
>> I think that there are 3 distinct uses:
>>=20
>> 1=2E As a light formatting/checking help, for instance to generate a ta=
ble=20
>from=20
>> a human written section, to correct the formulation of a sentence, or t=
o=20
>draft=20
>> an example=2E This is notably useful for non native English speakers=2E
>>=20
>> 2=2E As a help to search existing state of art on some feature, or any =
kind of=20
>> data, without writing anything in a protoXEP=2E
>>=20
>> 3=2E As a way to generate whole sections=2E
>>=20
>> Instinctively, and If we put aside ethical and ecological concerns abou=
t=20
>LLMs,=20
>> I think that 1=2E and 2=2E are OK, and 3=2E should be forbidden=2E And =
in all cases,=20
>> it should be disclosed=2E
>>=20
>> I would like your feedback on this matter, in particular people with le=
gal=20
>> knowledge=2E
>>=20
>> I would like to avoid a flamewar, I know that this topic is sensitive a=
nd=20
>there=20
>> opinions are highly divided, please express your opinion calmly=2E The =
fact=20
>is,=20
>> we can't ignore this anymore=2E
>>=20
>> Should this be discussed with board or council?
>>=20
>> Thanks=2E
>>=20
>> Best,
>> Goffi
>

------N6UEHOVTLDB10FEFDLE8LQDNZLIB5T
Content-Type: text/html;
 charset=utf-8
Content-Transfer-Encoding: quoted-printable

<html><head></head><body><div dir=3D"auto">Hi, <br><br>It is still not very=
 clear to me what the objective is and how a requirement to disclose the us=
e of AI achieves it=2E I see these angles: <br><br>1=2E AI may generate out=
put that is of low quality (incomplete, vague, false, superfluous)=2E<br><b=
r>2=2E AI may generate output that includes parts of other works without at=
tribution and/or permission (from a rights perspective)=2E<br><br>I do not =
see why these problems uniquely exist because of the use of AI, and why thi=
s isn't covered by an author's existing responsibility to ensure quality an=
d the pre-conditions for assigning ownership rights to the XSF per its IPR =
policy (in particular sections 3=2E1 and 3=2E2)=2E<br><br>If you cannot for=
mulate the requirements for quality for humans, then how does declaring use=
 of AI make things better? And if you can, and a submission complies, how d=
oes it matter that AI was used in the process? How would you handle =E2=80=
=9CAI-tainted=E2=80=9D submissions differently?<br><br>The same holds for p=
otential rights issues=2E The author is still responsible=2E <br><br>My con=
cern is that all future submissions will just have this disclosure as boile=
rplate, and a recipient cannot assess how deep the impact of the use of AI =
is=2E With the accelerating growth we see in both the capabilities and the =
use of AI, this becomes increasingly hard=2E We will not have gained anythi=
ng substantial=2E In that regard it reminds me of the Evil Bit (RFC 3514) /=
 Malicious Stanzas (XEP-0076)=2E<br><br>The statement you linked to seems l=
ike common sense and doesn't actually require disclosure=2E Consider this d=
ocument again, but conceptually replace =E2=80=9Cthe use of AI=E2=80=9D wit=
h the =E2=80=9Cuse of a keyboard=E2=80=9D=2E Does having such a statement c=
hange the outcome of our standards process?<br><br>I think that our author =
guidelines (XEP-0143) are already clear, and sense all the above boils down=
 to this:<br><br>=C2=A0=C2=A0=C2=A0 =E2=80=9CAuthors must understand, verif=
y, and take responsibility for every contribution they submit=2E=E2=80=9D (=
-- ChatGPT with my prompting)<br><br>While noting that this stance is not u=
niversally accepted (e=2Eg=2E see &lt;<a href=3D"https://jme=2Ebmj=2Ecom/co=
ntent/51/4/230">https://jme=2Ebmj=2Ecom/content/51/4/230</a>&gt;), I think =
it is suitable for our standards process=2E Feel free to use this phrase in=
 a concrete proposal, like a PR to XEP-0143=2E<br><br>Cheers, <br><br>ralph=
m<br><br>Disclosure: this message was manually (glide) typed on a OnePlus 1=
2, using Google Board, in Thunderbird for Android=2E Conversing with AI may=
 have influenced my thought process=2E Except where explicitly noted, no ex=
cerpts of other works were included in this message=2E<br><br></div><br><br=
><div class=3D"gmail_quote"><div dir=3D"auto">On 7 July 2026 17:55:54 CEST,=
 Goffi &lt;goffi@goffi=2Eorg&gt; wrote:</div><blockquote class=3D"gmail_quo=
te" style=3D"margin: 0pt 0pt 0pt 0=2E8ex; border-left: 1px solid rgb(204, 2=
04, 204); padding-left: 1ex;">
<pre class=3D"net-thunderbird-android-beta__plain-text-message-pre"><div d=
ir=3D"auto">Hi,<br><br>This discussion is stalling=2E<br><br>Meanwhile, we'=
ve got specification(s?) which have been clearly written with AI, <br>and I=
 guess it will be more and more often the case=2E<br><br>As a reviewer with=
 my council hat, I would really love to at least have <br>disclaimer when A=
I is used (is it for writing whole sections, to extract a <br>table, to wri=
te example, to check spelling/grammar)=2E<br><br>Many big projects have a s=
tatement on AI use, e=2Eg=2E, CPython:<br><a href=3D"https://devguide=2Epyt=
hon=2Eorg/getting-started/ai-tools/">https://devguide=2Epython=2Eorg/gettin=
g-started/ai-tools/</a><br><br>I think XSF should have one too=2E<br><br>Do=
 we need more discussion on standard, or should the board discuss that and =
<br>ask a team to work on a AI statement?<br><br>Thanks<br>Goffi<br><br><br=
>Le lundi 11 mai 2026, 09:50:19 heure d=E2=80=99=C3=A9t=C3=A9 d=E2=80=99Eur=
ope centrale Goffi a =C3=A9crit :<br></div><blockquote class=3D"gmail_quote=
" style=3D"margin-bottom: 1ex; --net-thunderbird-android-beta__blockquote-d=
efault-border-color: #729fcf;"><div dir=3D"auto">Hello everybody,<br><br>I =
would like to bring a discussion on AI policy=2E We can't really ignore <br=
>anymore that modern models have become very capable, and I suspect that th=
ey <br>are used for spec authoring=2E<br><br>This raises, I believe, copyri=
ght issues: if someone use AI to redact a <br></div></blockquote><div dir=
=3D"auto">whole <br></div><blockquote class=3D"gmail_quote" style=3D"margin=
-bottom: 1ex; --net-thunderbird-android-beta__blockquote-default-border-col=
or: #729fcf;"><div dir=3D"auto">section of a spec, how can we be sure that =
it's not an existing specs for <br></div></blockquote><div dir=3D"auto">som=
e <br></div><blockquote class=3D"gmail_quote" style=3D"margin-bottom: 1ex; =
--net-thunderbird-android-beta__blockquote-default-border-color: #729fcf;">=
<div dir=3D"auto">other place, possibly under copyright, that is copied or =
paraphrased? How <br></div></blockquote><div dir=3D"auto">can <br></div><bl=
ockquote class=3D"gmail_quote" style=3D"margin-bottom: 1ex; --net-thunderbi=
rd-android-beta__blockquote-default-border-color: #729fcf;"><div dir=3D"aut=
o">an author guarantee that it's original work (hint: they can't)?<br><br>I=
 think that there are 3 distinct uses:<br><br>1=2E As a light formatting/ch=
ecking help, for instance to generate a table <br></div></blockquote><div d=
ir=3D"auto">from <br></div><blockquote class=3D"gmail_quote" style=3D"margi=
n-bottom: 1ex; --net-thunderbird-android-beta__blockquote-default-border-co=
lor: #729fcf;"><div dir=3D"auto">a human written section, to correct the fo=
rmulation of a sentence, or to <br></div></blockquote><div dir=3D"auto">dra=
ft <br></div><blockquote class=3D"gmail_quote" style=3D"margin-bottom: 1ex;=
 --net-thunderbird-android-beta__blockquote-default-border-color: #729fcf;"=
><div dir=3D"auto">an example=2E This is notably useful for non native Engl=
ish speakers=2E<br><br>2=2E As a help to search existing state of art on so=
me feature, or any kind of <br>data, without writing anything in a protoXEP=
=2E<br><br>3=2E As a way to generate whole sections=2E<br><br>Instinctively=
, and If we put aside ethical and ecological concerns about <br></div></blo=
ckquote><div dir=3D"auto">LLMs, <br></div><blockquote class=3D"gmail_quote"=
 style=3D"margin-bottom: 1ex; --net-thunderbird-android-beta__blockquote-de=
fault-border-color: #729fcf;"><div dir=3D"auto">I think that 1=2E and 2=2E =
are OK, and 3=2E should be forbidden=2E And in all cases, <br>it should be =
disclosed=2E<br><br>I would like your feedback on this matter, in particula=
r people with legal <br>knowledge=2E<br><br>I would like to avoid a flamewa=
r, I know that this topic is sensitive and <br></div></blockquote><div dir=
=3D"auto">there <br></div><blockquote class=3D"gmail_quote" style=3D"margin=
-bottom: 1ex; --net-thunderbird-android-beta__blockquote-default-border-col=
or: #729fcf;"><div dir=3D"auto">opinions are highly divided, please express=
 your opinion calmly=2E The fact <br></div></blockquote><div dir=3D"auto">i=
s, <br></div><blockquote class=3D"gmail_quote" style=3D"margin-bottom: 1ex;=
 --net-thunderbird-android-beta__blockquote-default-border-color: #729fcf;"=
><div dir=3D"auto">we can't ignore this anymore=2E<br><br>Should this be di=
scussed with board or council?<br><br>Thanks=2E<br><br>Best,<br>Goffi<br></=
div></blockquote><div dir=3D"auto"><br></div></pre></blockquote></div></bod=
y></html>
------N6UEHOVTLDB10FEFDLE8LQDNZLIB5T--

--===============7752431507778712303==
Content-Type: text/plain; charset="us-ascii"
MIME-Version: 1.0
Content-Transfer-Encoding: 7bit
Content-Disposition: inline

_______________________________________________
Standards mailing list -- [email protected]
To unsubscribe send an email to [email protected]

--===============7752431507778712303==--