[rpd] Writing tools

[Mike Burns]

Hi Nonjabulo,

 

You seem to continue to miss the context of my comment about "Denial of
Service sort of attack" which was explicitly framed in the context of
swamping the list.

I also mentioned "But unless there is an attempt by an individual to swamp
the list, the posts should be treated as if written by the poster, that is
like every other post."

 

And I think there has been support for this separation of the "AI" issue
into more digestible pieces.

 

I don't want AI to be used as a tool to generate so many copies of the same
argument in slightly different words WITH AN INTENT TO SWAMP THE LIST, and
we know it is quite capable of that.

Why not a simple +1 comment if you want your argument noted and it has
already been stated by another?

 

However, as I have also said, I have not seen any deployment of AI in this
manner in the current discussions, and my comment was broader in target.

Sorry for any confusion and I hope I have answered your concern. 

 

I don't want AI generated messages to overwhelm any policy list, but I don't
care if list members employ AI as a writing tool.

In some ways it could be an improvement.

 

Regards,
Mike

 

 

From: Nonjabulo Sphilile <nonjabulosphilile at gmail.com> 
Sent: Monday, July 20, 2026 1:36 PM
To: Mike Burns <mike at iptrading.com>
Cc: rpd at afrinic.net
Subject: Re: [rpd] Writing tools

 

Hi Mike,

 

Your clarification does not remove my concern. Calling AI-assisted
participation a potential "DDoS" still frames the tool as suspicious before
any disruptive conduct has occurred.

 

Existing rules already address flooding. Moderation should respond to proven
behaviour, not hypothetical assumptions about technology. Otherwise, process
management becomes gatekeeping.

 

Regards,

Nonjabulo

 

On Mon, 20 Jul 2026, 18:14 Mike Burns, <mike at iptrading.com
<mailto:mike at iptrading.com> > wrote:

Hi Nonjablulo,

 

I want to be clear that AI DDos is the same as automated flooding. AI can
generate non-duplicative prompts more quickly and easily than anything else,
which provide a tool for list-swamping.

What that can do is render the list unusable. 

Note that I do not see that happening in any current conversations.

And as I pointed out, existing procedures can take care of list flooding,
whether AI generated or not.

 

I thought I was clear that absent the flooding I called a DDoS, all
arguments should be considered on their merits.

I personally don't care where the argument comes from.

 

Regards,
Mike

 

 

From: Nonjabulo Sphilile < <mailto:nonjabulosphilile at gmail.com>
nonjabulosphilile at gmail.com> 
Sent: Monday, July 20, 2026 12:02 PM
To:  <mailto:rpd at afrinic.net> rpd at afrinic.net
Subject: Re: [rpd] Writing tools

 

Dear Mike,

 

The "AI DDoS" analogy is the wrong category. Participants are not packets,
and disagreement is not an attack on the process.

 

The list may regulate actual conduct such as automated flooding,
impersonation, excessive duplication, or deliberate disruption. It should
not infer abuse from writing style or from the use of a drafting tool. A
person who reviews, submits, and stands behind a message remains responsible
for it.

 

Any guideline must therefore be tool-neutral and based on measurable
behaviour. Otherwise process management becomes a mechanism for filtering
unfamiliar or inconvenient voices while established participants continue to
define what counts as legitimate participation.

 

Rough consensus is not protected by narrowing the room. It is protected by
identifying distinct issues, answering them on their merits, and refusing to
turn procedural control into institutional mandate.

 

I oppose any rule that treats AI assistance itself as presumptive
misconduct.

 

Regards,

Nonjabulo

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.afrinic.net/pipermail/rpd/attachments/20260720/fbcf20e8/attachment-0001.html>