• Re: Do we need a conflict of interest policy?

    From =?utf-8?Q?Pierre-Elliott_B=C3=A9cue@21:1/5 to G. Branden Robinson on Sat Feb 8 01:10:02 2025
    "G. Branden Robinson" <g.branden.robinson@gmail.com> wrote on 07/02/2025 at 17:41:41+0100:

    [M-F-T set to -project; see https://cr.yp.to/proto/replyto.html ]

    At 2025-02-07T17:08:44+0100, Pierre-Elliott Bécue wrote:
    I'm not sure I'd expect more from people than them saying who they
    work for or defend.

    I'd not, eg, expect some Canonical employee to refrain voting a GR
    because they might be biased. We all are biased.

    I think this is a non sequitur. Yes, we are all human (except for the
    LLM that wrote Charles's draft policy), and therefore we all have
    cognitive biases.

    That does _not_ imply that we all have conflicts of interest.

    Biases are often unconscious, whereas conflicts of interest are always
    known to the person who has them.

    I don't think it's always clear to them.

    But I get your point and I agree that there's a difference.

    I'm not sure that, considering that we are around a 1000 people here, it
    really is an issue.

    --
    PEB

    --=-=-Content-Type: application/pgp-signature; name="signature.asc"

    -----BEGIN PGP SIGNATURE-----

    iQJDBAEBCgAtFiEE5CQeth7uIW7ehIz87iFbn7jEWwsFAmemnXoPHHBlYkBkZWJp YW4ub3JnAAoJEO4hW5+4xFsL7YkQAJS/Wn795QwEC0bbHnZrSmCYrX57nqE5FhCd qxMdKFzS9lyRYXSza+UL7bUiQUgKI0OLXw0N/9noDLdXsuAgwL5pCMUoNk6HmidJ LQs6RBmPiBpG5jrkmaulSvWKKgaAKMF4JB9VBa40VlFsd01ydsQ+lxWfsLrIDg3m n6sPMIzEkyHuMUEjr5sMRAuaqrGsiMW0LT9/dHnbGsmABAhqI5fa09lziMGm0t8u py/7JD74Gbyaa24A0ep6u27133p2YRgwKkZOur27vifrjhkcGbFtaz3NUYjBWMCK lc9xlT8B2ln1G55xxeZsMEymEeZzrxiZU5q/9ewXak8R0R7peajw0IvK80S9ORIp XY8eM4ZnItx+rQJewqqGzP05yyxZOFCnlL8eNIQMEsk9+2Un1UrzonmG68/AKy6D HIRwRYIybpa1euMBv+Bj8Si7/VvpbgoP5SqpOvtm+suKi+QxpZPZ/56l5QpHxlAc QOtphwERBAJgQJPD1xXflzu6rqctKBkFOraW2Kvd9g7NQGcVvggzkvIz5LT/1+fP 17WWfMewsBXcQCR4Db7BHjDrN2NF5oEdUCax75oxRV6Iio6632TYlBke+4WerAsm HM8BFYzy/BU49UFg0bUd7DB9UfRqrwKrQarMBRsWwuZZKSIvb+jHeaOFCB8nUO55
    my9/QFov
    =RaNp
    -----END PGP SIGNATURE-----

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Charles Plessy@21:1/5 to All on Sat Feb 8 02:10:01 2025
    "G. Branden Robinson" <g.branden.robinson@gmail.com> wrote on 07/02/2025 at 17:41:41+0100:
    Biases are often unconscious, whereas conflicts of interest are always known to the person who has them.

    Le Sat, Feb 08, 2025 at 12:55:38AM +0100, Pierre-Elliott BΘcue a Θcrit :
    I don't think it's always clear to them.

    At my workplace, we have "rules for conflicts of interest management",
    and in of them aims at "preventing situations in which [the researchers]
    may be perceived to have a conflict of interest".

    So here we have three different scopes:

    - conflict of interest is concious to individuals,
    - conflict of interest can be unconcious to individuals,
    - conflict of interest can be outside the control of individuals.

    As we see in this discussion, there is a considerable disagreement
    within Debian about the very definition of what a conflict of interest
    is. I think that this makes us ill-equipped to to handle situations
    where one developer expresses the opinion that another developer has a
    conflict of interest.

    We are not going to solve that overnight, but I feel that we need to
    agree on the minimum, which is:

    - it is not a taboo,
    - it is not an attack,
    - it is not a misconduct.

    This is all that I have been aiming for when I started this discussion.

    By the way, for the full disclosure, and also to counter the argument
    that using ChatGPT to write a text in a language that I am not
    comfortable with was an act of carelessness in which I dumped
    low-quality text produced with no effort, here is the inputs I gave:

    Prompt 1:

    Please draft a very brief conflict of interest statement for a
    non-profit organisation such as Debian, that has a very diverse
    membership, and diverse understanding, acceptation and experience of the concept of conflict of interest. The main point to address is that
    people have employers or sponsors which themelves can be donators, have interest, or be impacted by Debian's decisions. The policy should advise
    on how to handle the information and when to refrain from participating
    to a decision, and also how to raise the possibility of a conflict of
    interest without offending the person, etc.

    Not satisfied with the results, I added prompt 2:

    Can it explain a bit more what a conflict of interest is, acknowledge
    that opinions may differ on that point, and also say that there is
    nothing bad at all of being in a situation of conflict of interest?

    I was then satisfied that what I read was what I wanted to write, but
    better written that if I would do it myself.

    Some parts are not perfect, but if there is appetite we have a lot of
    talented people here who can make it better.

    We do not have to conclude soon. We can give the chance to everybody, including the vast majority of people who read but do not write on this
    list, to think about it, take the time it needs to be exposed to
    situations in their life where the question of "what is a conflict of
    interest" has some relevance, before we try to progress further.

    Have a nice week-end,

    Charles

    --
    Charles Plessy Nagahama, Yomitan, Okinawa, Japan
    Debian Med packaging team http://www.debian.org/devel/debian-med Tooting from home https://framapiaf.org/@charles_plessy
    - You do not have my permission to use this email to train an AI -

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Timo =?utf-8?Q?R=C3=B6hling?=@21:1/5 to All on Sat Feb 8 10:00:01 2025
    Hi Charles,

    * Charles Plessy <plessy@debian.org> [2025-02-08 10:02]:
    We are not going to solve that overnight, but I feel that we need
    to agree on the minimum, which is:

    - it is not a taboo,
    - it is not an attack,
    - it is not a misconduct.

    +1

    By the way, for the full disclosure, and also to counter the
    argument that using ChatGPT to write a text in a language that I am
    not comfortable with was an act of carelessness in which I dumped >low-quality text produced with no effort, here is the inputs I
    gave:
    [...]
    For the record and notwithstanding my humorous take on model
    collapse, I happen to think that it is absolutely okay to use LLMs
    like this, because their primary function is text generation, and
    they do this very well. Arguably better than me, at the very least.

    And let's not forget we are talking about a draft here, reviewed and
    refined by someone with domain knowledge at that. This is not
    someone replacing critical Linux kernel code with untested Copilot
    drivel, nor an attempt to cheat an exam or fake a scientific
    publication.

    Have a nice week-end,
    Likewise!

    Cheers
    Timo

    --
    ⢀⣴⠾⠻⢶⣦⠀ ╭────────────────────────────────────────────────────╮
    ⣾⠁⢠⠒⠀⣿⡁ │ Timo Röhling │
    ⢿⡄⠘⠷⠚⠋⠀ │ 9B03 EBB9 8300 DF97 C2B1 23BF CC8C 6BDD 1403 F4CA │
    ⠈⠳⣄⠀⠀⠀⠀ ╰────────────────────────────────────────────────────╯

    -----BEGIN PGP SIGNATURE-----

    iQIzBAEBCgAdFiEEmwPruYMA35fCsSO/zIxr3RQD9MoFAmenFIAACgkQzIxr3RQD 9Mq9fRAAgbpQrF6zfsRvAlbRVZq4gS6cbZOcNtQuYFHfCbeRhbg4LkmSS5UR0oLT TTHuVRNYyp6lyj34CxXVZVtOJmveYaSVV0zPOdvpXNk9c1/1r2hNnjjng1n6S08B gdhEa5T6LtAd+8XbTS1PfztUxc5UAM6msUn1+2mXb2Rm2M2Fn9ThLpxnvgAUU3PQ etlDiKFGNTlLLHhfme6T1XStSKEcjGkz6LR4EiQ0FKVNLhw2YUhUOP+VvKbgSQse redvdFrjCm/FmmG1K3WMHRa1KKYMZeXIB2ZFtRtwedPn90KVaSEQbtgUAy0g/j4o Fz/Lb3+RJ6OlCjtblDCzshBWXy/xsn184Pgg/Cmmxpr
  • From Jonathan Dowland@21:1/5 to All on Thu Mar 6 12:20:01 2025
    I stumbled across this Debian Wiki page I'd forgotten about. We have had
    a voluntary Register of Interests here since 2017:

    https://wiki.debian.org/RegisterOfInterests


    --
    Please do not CC me for listmail.

    👱🏻 Jonathan Dowland
    jmtd@debian.org
    🔗 https://jmtd.net

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)