Re: [NNagain] The growing challenges of discerning authentic vs. inauthentic information and identity

[Jack Haverty <jack@3kitty.org>]

[-- Attachment #1.1.1.1: Type: text/plain, Size: 4075 bytes --]

IMHO, similar issues of judgement and trust have come up in the past, 
and it might be worth researching the history.

In the context of the Web, during the 90s there was a similar concern 
about categorizing content available on the Internet.   The issue at the 
time was providing mechanisms to protect children from pornography.  But 
today's issues of truth and misinformation are very similar -- e.g., you 
might categorize an inaccurate news post as "pornographic".

I suggest looking at some work from the 90s.  At the time, I was working 
at Oracle as "Internet Architect", and served as corporate 
representative to W3C (see https://www.w3.org/ ).  Thw W3C group, led by 
Tim Berners-Lee, was intensely involved in setting technical standards 
for the Web.

A project was formed call PICS - Platform for Internet Content 
Selection.  Essentially it created mechanisms to add metadata to 
existing content on the Web, and use it to filter content for end users.

See https://www.w3.org/PICS/ for the history.  PICS is now obsolete and 
was replaced by something called POWDER - see 
https://www.w3.org/2007/powder/

I wasn't involved in POWDER, which occurred after my involvement with 
W3C ended.  But I was very involved in the creation of PICS.

The main idea of PICS was to enable the creation of "rating schemes" to 
categorize content.  Since the focus was on pornography, one likely 
rating scheme was the classical G/R/X ratings popular at the time for 
characterizing movies.   But anyone, or any group, could define a rating 
scheme to suit their views.

Having selected a rating scheme they liked, any group, or individual, 
could assign ratings to specific content.  Perhaps you think that movie 
is "R", but I think it's "X".   As a judge once noted - "I can't define 
it, but I know it when I see it".   Opinions can of course differ.

Ratings were to be kept in one or more databases, accessible on the 
Internet to anyone.  Content could be identified by a URL, or perhaps a 
unique cryptographic "hash" of the content itself, in case it was 
moved.  Each record would contain 4 items - the identity of the content, 
the identity of the rating scheme used, the identity of the person or 
group making the rating, and the rating which they assigned.  Such 
technology was easily within the capabilities of databases even then.

On the "consumer" side, applications (e.g., browsers) would have 
settings that could be applied to indicate which rating system was to be 
used, which groups or persons making ratings were to be trusted, and 
what ratings of content would be actually viewable by the human end user.

The idea was that various groups (content creators, reviewers, religious 
groups, community activists, etc.) would define their preferred rating 
scheme and then assign ratings, at least to content they deemed 
objectionable.

End users, e.g., parents, could then set up their children's web 
browsers to use the rating scheme of whichever group(s) they trusted to 
make "correct" ratings, and set their children's browsers appropriately 
to restrict the content they could see.   A content consumer simply 
selects the rating service they trust.

It seems straightforward how a similar mechanism might be applied to 
instead rate accuracy of Internet content, and allow consumers to choose 
which, if any, ratings are applied to filter the information they see, 
based on who they trust to make such judgements.

PICS was actually implemented in popular browser software.   But, as far 
as I know, no group ever designed their preferred rating scheme, or 
actually assigned ratings to any content then available on the 
Internet.   The mechanisms were there.   But apparently no one used 
them.   The loud voices of "Something has to be done!" didn't actually 
themselves do anything.

Even if PICS/POWDER isn't appropriate for handling misinformation, an 
analysis of why it failed to be used might be revealing.

Jack Haverty



[-- Attachment #1.1.1.2: Type: text/html, Size: 4865 bytes --]

[-- Attachment #1.1.2: OpenPGP public key --]
[-- Type: application/pgp-keys, Size: 2469 bytes --]

[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 665 bytes --]