[NNagain] Introduction: Dr. David Bray
Jack Haverty
jack at 3kitty.org
Sat Oct 7 16:10:00 EDT 2023
Hi again David et al,
Interesting frenzy...lots of questions that need answers and associated
policies. I served 6 years as an elected official (in a small special
district in California), so I have some small understanding of the
government side of things and the constraints involved. Being in
charge doesn't mean you can do what you want.
I'm thinking here more near-term and incremental steps. You said "These
same questions need pragmatic pilots that involve the public ..."
So, how about using the current NN situation for a pilot? Keep all the
current ways and emerging AI techniques to continue to flood the system
with comments. But also offer an *optional* way for humans to
"register" as a commenter and then submit their (latest only) comment
into the melee. Will people use it? Will "consumers" (the lawyers,
commissioners, etc.) find it useful?
I've found it curious, for decades now, that there are (too many)
mechanisms for "secure email", that may help with the flood of
disinformation from anonymous senders, but very very few people use
them. Maybe they don't know how; maybe the available schemes are too
flawed; maybe ...?
About 30 years ago, I was a speaker in a public meeting orchestrated by
USPS, and recommended that they take a lead role, e.g., by acting as a
national CA - certificate authority. Never happened though. FCC issues
lots of licenses...perhaps they could issue online credentials too?
Perhaps a "pilot" where you will also accept comments by email, some
possibly sent by "verified" humans if they understand how to do so,
would be worth trying? Perhaps comments on "technical aspects" coming
from people who demonstrably know how to use technology would be
valuable to the policy makers?
The Internet, and technology such as TCP, began as an experimental pilot
about 50 years ago. Sometimes pilots become infrastructures.
FYI, I'm signing this message. Using OpenPGP. I could encrypt it also,
but my email program can't find your public key.
Jack Haverty
On 10/5/23 14:21, David Bray, PhD wrote:
> Indeed Jack - a few things to balance - the Administrative Procedure
> Act of 1946 (on which the idea of rulemaking is based) us about
> raising legal concerns that must be answered by the agency at the time
> the rulemaking is done. It's not a vote nor is it the case that if the
> agency gets tons of comments in one direction that they have to go in
> that direction. Instead it's only about making sure legal concerns are
> considered and responded to before the agency before the agency acts.
> (Which is partly why sending "I'm for XYZ" or "I'm against ABC" really
> doesn't mean anything to an agency - not only is that not a legal
> argument or concern, it's also not something where they're obligated
> to follow these comments - it's not a vote or poll).
>
> That said, political folks have spun things to the public as if it is
> a poll/vote/chance to act. The raise a valid legal concern part of the
> APA of 1946 is omitted. Moreover the fact that third party law firms
> and others like to submit comments on behalf of clients - there will
> always be a third party submitting multiple comments for their clients
> (or "clients") because that's their business.
>
> In the lead up to 2017, the Consumer and Government Affairs Bureau of
> the FCC got an inquiry from a firm asking how they could submit 1
> million comments a day on an "upcoming privacy proceeding" (their
> words, astute observers will note there was no privacy proceeding
> before the FCC in 2017). When the Bureau asked me, I told them either
> mail us a CD to upload it or submit one comment with 1 million
> signatures. To attempt to flood us with 1 million comments a day
> (aside from the fact who can "predict" having that many daily) would
> deny resources to others. In the mess that followed, what was released
> to the public was so redacted you couldn't see the legitimate concerns
> and better paths that were offered to this entity.
>
> And the FCC isn't alone. EPA, FTC, and other regulatory agencies have
> had these hijinks for years - and before the Internet it was faxes,
> mass mimeographs (remember blue ink?), and postcards.The
> Administrative Conference of the United States (ACUS) - is the body
> that is supposed to provide consistent guidance for things like this
> across the U.S. government. I've briefed them and tried to raise
> awareness of these issues - as I think fundamentally this is a
> **process** question that once answered, tech can support. However
> they're not technologies and updating the interpretation of the
> process isn't something lawyers are apt to do until the evidence that
> things are in trouble is overwhelming.
>
> 52 folks wrote a letter to them - and to GSA - back in 2020. GSA had a
> rulemaking of its own on how to improve things, yet oddly never
> published any of the comments it received (including ours) and closed
> the rulemaking quietly. Here's the letter:
> https://tinyurl.com/letter-signed-52-people
>
> And here's an article published in OODAloop about this - and why
> Generative AI is probably going to make things even more challenging:
> https://www.oodaloop.com/archive/2023/04/18/why-a-pause-on-ai-development-is-not-the-answer-an-insiders-perspective/
>
> [snippet of the article] *Now in 2023 and Beyond: Proactive Approaches
> to AI and Society*
>
> Looking to the future, to effectively address the challenges arising
> from AI, we must foster a proactive, results-oriented, and cooperative
> approach with the public
> <https://davidbray.medium.com/challenges-and-needed-new-solutions-for-open-societies-to-maintain-civil-discourse-part-1-b5ea95f8c679>.
> Think tanks and universities can engage the public in conversations
> about how to work, live, govern, and co-exist with modern technologies
> that impact society. By involving diverse voices in the
> decision-making process, we can better address and resolve the complex
> challenges AI presents on local and national levels.
>
> In addition, we must encourage industry and political leaders to
> participate in finding non-partisan, multi-sector solutions if civil
> societies are to remain stable. By working together, we can bridge the
> gap between technological advancements and their societal implications.
>
> Finally, launching AI pilots across various sectors, such as work,
> education, health, law, and civil society, is essential. We must learn
> by doing on how we can create responsible civil environments where AIs
> can be developed and deployed responsibly. These initiatives can help
> us better understand and integrate AI into our lives, ensuring its
> potential is harnessed for the greater good while mitigating risks.
>
> In 2019 and 2020, a group of fifty-two people asked the Administrative
> Conference of the United States
> <https://tinyurl.com/letter-signed-52-people>(which helps guide
> rulemaking procedures for federal agencies), General Accounting
> Office, and the General Services Administration to call attention to
> the need to address the challenges of chatbots flooding public
> commenting procedures and potentially crowding out or denying services
> to actual humans wanting to leave a comment. We asked
> <https://davidbray.medium.com/challenges-and-needed-new-solutions-for-open-societies-to-maintain-civil-discourse-part-1-b5ea95f8c679>:
>
>
> *1. Does identity matter regarding who files a comment or not — and
> must one be a U.S. person in order to file?*
>
> *2. Should agencies publish real-time counts of the number of comments
> received — or is it better to wait until the end of a commenting round
> to make all comments available, including counts?*
>
> *3. Should third-party groups be able to file on behalf of someone
> else or not — and do agencies have the right to remove spam-like
> comments?*
>
> *4. Should the public commenting process permit multiple comments per
> individual for a proceeding — and if so, how many comments from a
> single individual are too many? 100? 1000? More?*
>
> *5. Finally, should the U.S. government itself consider, given public
> perceptions about potential conflicts of interest for any agency
> performing a public commenting process, whether it would be better to
> have third-party groups take responsibility for assembling comments
> and then filing those comments via a validated process with the
> government?*
>
> These same questions need pragmatic pilots that involve the public to
> co-explore and co-develop how we operate effectively amid these
> technological shifts
> <https://davidbray.medium.com/challenges-and-needed-new-solutions-for-open-societies-to-maintain-civil-discourse-part-2-2f637c472112>.
> As the capabilities of LLMs continue to grow, we need positive change
> agents willing to tackle the messy issues at the intersection of
> technology and society. The challenges are immense, but so too are the
> opportunities for positive change. Let’s seize this moment to create a
> better tomorrow for all. Working together, we can co-create a future
> that embraces AI’s potential while mitigating its risks
> <https://medium.com/peoplecentered/the-need-for-people-centered-sources-of-hope-for-our-digital-future-ahead-ef491dd2703d>,
> informed by the hard lessons we have already learned.
>
> Full article:
> https://www.oodaloop.com/archive/2023/04/18/why-a-pause-on-ai-development-is-not-the-answer-an-insiders-perspective/
>
> Hope this helps.
>
>
> On Thu, Oct 5, 2023 at 4:44 PM Jack Haverty via Nnagain
> <nnagain at lists.bufferbloat.net> wrote:
>
> Thanks for all your efforts to keep the "feedback loop" to the
> rulemakers functioning!
>
> I'd like to offer a suggestion for a hopefully politically
> acceptable way to handle the deluge, derived from my own battles
> with "email" over the years (decades).
>
> Back in the 1970s, I implemented one of the first email systems on
> the Arpanet, under the mentorship of JCR Licklider, who had been
> pursuing his vision of a "Galactic Network" at ARPA and MIT. One
> of the things we discovered was the significance of anonymity.
> At the time, anonymity was forbidden on the Arpanet; you needed an
> account on some computer, protected by passwords, in order to
> legitimately use the network. The mechanisms were crude and
> easily broken, but the principle applied.
>
> Over the years, that principle has been forgotten, and the right
> to be anonymous has become entrenched. But many uses of the
> network, and needs of its users, demand accountability, so all
> sorts of mechanisms have been pasted on top of the network to
> provide ways to judge user identity. Banks, medical services,
> governments, and businesses all demand some way of proving your
> identity, with passwords, various schemes of 2FA, VPNs, or other
> such technology, with varying degrees of protection. It is still
> possible to be anonymous on the net, but many things you do
> require you to prove, to some extent, who you are.
>
> So, my suggestion for handling the deluge of "comments" is:
>
> 1/ create some mechanism for "registering" your intent to submit a
> comment. Make it hard for bots to register. Perhaps you can
> leverage the work of various partners, e.g., ISPs, retailers,
> government agencies, financial institutions, of others who already
> have some way of identifying their users.
>
> 2/ Also make registration optional - anyone can still submit
> comments anonymously if they choose.
>
> 3/ for "registered commenters", provide a way to "edit" your
> previous comment - i.e., advise that your comment is always the
> last one you submitted. I.E., whoever you are, you can only
> submit one comment, which will be the last one you submit.
>
> 4/ In the thousands of pages of comments, somehow flag the ones
> that are from registered commenters, visible to the people who
> read the comments. Even better, provide those "information
> consumers" with ways to sort, filter, and search through the body
> of comments.
>
> This may not reduce the deluge of comments, but I'd expect it to
> help the lawyers and politicians keep their heads above the water.
>
> Anonymity is an important issue for Net Neutrality too, but I'll
> opine about that separately.....
>
> Jack Haverty
>
>
> On 10/2/23 12:38, David Bray, PhD via Nnagain wrote:
>> Greetings all and thank you Dave Taht for that very kind intro...
>>
>> First, I'll open with I'm a gosh-darn non-partisan, which means I
>> swore an oath to uphold the Constitution first and serve the
>> United States - not a specific party, tribe, or ideology. This
>> often means, especially in today's era of 24/7 news and social
>> media, non-partisans have to "top cover".
>>
>> Second, I'll share that in what happened in 2017 (which itself
>> was 10x what we saw in 2014) my biggest concern was and remains
>> that a few actors attempted to flood the system with
>> less-than-authentic comments.
>>
>> In some respects this is not new. The whole "notice and comment"
>> process is a legacy process that goes back decades. And the FCC
>> (and others) have had postcard floods of comments, mimeographed
>> letters of comments, faxed floods of comments, and now this -
>> which, when combined with generative AI, will be yet another flood.
>>
>> Which gets me to my biggest concern as a non-partisan in
>> 2023-2024, namely how LLMs might misuse and abuse the commenting
>> process further.
>>
>> Both in 2014 and 2017, I asked FCC General Counsel if I could use
>> CAPTChA to try to reduce the volume of web scrapers or bots both
>> filing and pulling info from the Electronic Comment Filing System.
>>
>> Both times I was told *no* out of concerns that they might
>> prevent someone from filing. I asked if I could block obvious
>> spam, defined as someone filing a comment >100 times a minute,
>> and was similarly told no because one of those possible comments
>> might be genuine and/or it could be an ex party filing en masse
>> for others.
>>
>> For 2017 we had to spin up 30x the number of AWS cloud instances
>> to handle the load - and this was a flood of comments at 4am,
>> 5am, and 6am ET at night which normally shouldn’t see such
>> volumes. When I said there was a combination of actual humans
>> wanting to leave comments and others who were effectively denying
>> service to others (especially because if anyone wanted to do a
>> batch upload of 100,000 comments or more they could submit a CSV
>> file or a comment with 100,000 signatories) - both parties said
>> no, that couldn’t be happening.
>>
>> Until 2021 when the NY Attorney General proved that was exactly
>> what was happening with 18m of the 23m apparently from
>> non-authentic origin with ~9m from one side of the political
>> aisle (and six companies) and ~9m from the other side of the
>> political aisle (and one or more teenagers).
>>
>> So with Net Neutrality back on the agenda - here’s a simple
>> prediction, even if the volume of comments is somehow controlled,
>> 10,000+ pages of comments produced by ChatGPT or a different LLM
>> is both possible and probably will be done. The question is if
>> someone includes a legitimate legal argument on page 6,517 - will
>> FCC’s lawyers spot it and respond to it as part of the NPRM?
>>
>> Hope this helps and with highest regards,
>>
>> -d.
>> --
>>
>> Principal, LeadDoAdapt Ventures, Inc.
>> <https://www.leaddoadapt.com/> & Distinguished Fellow
>>
>> Henry S. Stimson Center
>> <https://www.stimson.org/ppl/david-bray/>, Business Executives
>> for National Security <https://bens.org/people/dr-david-bray/>
>>
>>
>>
>> On Mon, Oct 2, 2023 at 2:15 PM Dave Taht via Nnagain
>> <nnagain at lists.bufferbloat.net> wrote:
>>
>> All:
>>
>> I have spent the last several days reaching out to as many
>> people I
>> know with a deep understanding of the policy and technical issues
>> surrounding the internet, to participate on this list. I
>> encourage you
>> all to reach out on your own, especially to those that you can
>> constructively and civilly disagree with, and hopefully work
>> with, to
>> establish technical steps forward. Quite a few have joined
>> silently!
>> So far, 168 people have joined!
>>
>> Please welcome Dr David Bray[1], a self-described "human
>> flack jacket"
>> who, in the last NN debate, stood up for the non -partisan
>> FCC IT team
>> that successfully kept the system up 99.4% of the time
>> despite the
>> comment floods and network abuses from all sides. He has
>> shared with
>> me privately many sad (and some hilarious!) stories of that
>> era, and I
>> do kind of hope now, that some of that history surfaces, and
>> we can
>> learn from it.
>>
>> Thank you very much, David, for putting down your painful
>> memories[2],
>> and agreeing to join here. There is a lot to tackle here, going
>> forward.
>>
>> [1] https://www.stimson.org/ppl/david-bray/
>> [2] "Pain shared is reduced. Joy shared, increased." - Spider
>> Robinson
>>
>>
>> --
>> Oct 30:
>> https://netdevconf.info/0x17/news/the-maestro-and-the-music-bof.html
>> Dave Täht CSO, LibreQos
>> _______________________________________________
>> Nnagain mailing list
>> Nnagain at lists.bufferbloat.net
>> https://lists.bufferbloat.net/listinfo/nnagain
>>
>>
>> _______________________________________________
>> Nnagain mailing list
>> Nnagain at lists.bufferbloat.net
>> https://lists.bufferbloat.net/listinfo/nnagain
>
> _______________________________________________
> Nnagain mailing list
> Nnagain at lists.bufferbloat.net
> https://lists.bufferbloat.net/listinfo/nnagain
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.bufferbloat.net/pipermail/nnagain/attachments/20231007/d57cd6d2/attachment-0001.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: OpenPGP_signature.asc
Type: application/pgp-signature
Size: 665 bytes
Desc: OpenPGP digital signature
URL: <https://lists.bufferbloat.net/pipermail/nnagain/attachments/20231007/d57cd6d2/attachment-0001.sig>
More information about the Nnagain
mailing list