The proof that newspapers’ reader-commenters are data-collection gold is in site-owners’ tightening restrictions; attempts to control the historical record; and seizing gifts to the commons for themselves
The tiger mural is a reminder that our most ordinary actions, from shopping to pumping fuel ( above and below ) are being included in data-gathering for AI everywhere. In an old Chinese philosophical text, tigers spy with ‘sharp eyes’ — the official name of China’s nationwide surveillance system…
…
…
Hello froggy.
Happy New Year!
Before I get to frog ponds — or another way to think of comments sections of newspapers and other media sites — let’s think about where we are on this road to a world run on AI. What’s being done to commenters couldn’t be more emblematic.
No I’m not addressing you as a twitchy amphibian to be rude. I’m a frog too. We all are. Frogs clueless for the most part about the heat being turned up under this pot of water we’ve been put in ever so gently, as the dial twists by infinitesimal degrees. Or not so oblivious, but lying to ourselves about what is happening because truth is frightening when we don’t know how to defend ourselves against its implications.
What truth is that? you ask. This: in digital technology’s reshaping of society, we have gone far beyond the surveillance business model (or ‘surveillance capitalism’ if you are politically-minded) to losing our human agency — which is, being deprived of our ability to act on our intentions, or do freely what we reasonably expect to be able to. Or did.
Pay attention to what the most objective leading scientists are saying. Not the elite compromised by still working in or closely connected to technology giants, investors and startups. Not anyone taking Big Tech’s coin in research funding at their universities — the subject of a recent report in the Washington Post — unless they are as honest as the Belgian physicist Bob Coecke. He taught at Oxford for over twenty years and lately, has been doing all he can for emergency public education. In an interview with a Guardian journalist published two weeks ago he said, too bluntly for delicate ears:
‘Think about AI. Think about how the world is getting fucked up now. Billion-dollar companies are in charge of a revolution that could control the world and nobody understands what they are doing.’
Note his use of the present tense, dear froggy. Treat as singing with forked tongues the chorus of alarmism about AI dangers from too many sci-tech stars. Note that their warnings are focused on the future — with no mention of damage being done by AI in the present, in which fortunes are being made by some of those alarmists as they themselves exploit the all-invading technology (last mentioned in a post here in June.)
It’s hard to shake off the impression that the persistent future-focus is being used exactly like the ancient magician’s art of theatrical misdirection. A stream of deceptive patter from the conjurer concentrates on what he wants the audience to look at while he grabs the diversion to play his tricks and pull off his stunts.
How do you demonstrate that we are losing human agency? O froggy, o froggy, the proof that we’re frogs set for boiling is all around us in the trivia of daily life. It looks boring beside the futurist-conjurers’ high-flown speculations about AI becoming ‘conscious,’ and their arguments about what that what would mean.
You can trace an arc of horror from a New York Times exposé five years ago to a recent leak about Google’s objective in its nonstop data collection on, and analysis of us as individuals. In 2019 Kashmir Hill, a NYT reporter, discovered that Sifted, a company specialising in data-gathering, was a top purveyor of detailed personal profiles of consumers that other companies buy to decide on discriminatory treatment. How long you are kept on hold as you wait for customer support. Or when you have to return a defective product you have bought, make that task easy, arduous or impossible.
This is all done automatically by software, early AI or machine intelligence, based on your very personal Sifted profiling and categorisation, which can include — as the NYT journalist discovered — ten years of comments that you posted on some website. Rating like this could account for some cases in a Guardian report on 19 December about the frantic struggles of UK telecommunications customers attempting to cancel broadband contracts. One said that he could not get anyone to help him on his telecoms company’s web chat service ‘after leaving it open all day for two consecutive days,’ — after which he spent hours on hold seeking help by telephone, during which he was disconnected four times.
Who could expect any of this to get better after considering the implications of these Daily Mail gleanings from documents presented at a recent internal conference at Google describing ‘the tech giant’s plan to create an artificial intelligence designed to become its users’ “Life Story Teller”’ — software anyone could open for the first time and discover that ‘“it already knows everything about your life.’’’
As people paying attention know by now, dear froggy, the technologists say that they are only following digital technology where it is leading us unavoidably. Before the wide public debate that this rationalisation needs — urgently, members of the public have to open their eyes wide to their frogs-in-pot predicament.
…
My own enlightenment came from thinking about roughly a decade of reading and commenting on the website of the Financial Times, a newspaper to which I have had fond, work-related ties going back to three weeks before I turned twenty-one, a whole geological era ago. I mention this because feeling the temperature go up in the frog water on that site, and realising that commenter-frogs are no longer in a pond but a confining pot — thanks to sweeping changes in the online FT’s management practices — is especially unnerving for those of us who knew the paper in its still recent, pre-digital past.
There was of course no glimmer, in that past, of anything like the proposal related to this post-Gutenberg site’s launch in 2011 — a description of a new economic structure or ‘business model’ for publishing I called a keiretsu-cooperative, which placed reader-commenters at its core. I saw a chance for a sort of symphysis: commenters and publishers evolving, growing and prospering together — after an experimental phase in which the idea was debugged and refined.
In spite of the annoyance of trigger-happy site moderators, the first few years of media commenting sites, from approximately the mid-noughties, were glorious fun. Very happy frog ponds. On the websites where I tried out commenting — or unmediated publishing — it was startling to see readers hopping in and out at any hour of day or night, frequently posting links to their personal blogs and websites. We started our own, separate conversations in those places. Some of us went on to exchange news about our families as a matter of course.
The Financial Times site was not one of my early commenting ponds. Perhaps because it was already behind a paywall in 2016, and because a large proportion of its readers’ subscriptions are paid for by their expense accounts, the conversations there are less relaxed — if more reliably intelligent. But it was enough like the sites where I learnt how commenting works that I have been alarmed by registering signs of readers being treated over the years less like cherished visitors and more and more like data-collection objects (or dupes): instances adding up to froggy water warming up steadily but barely noticeably.
Signs such as these:
— Blocking or interfering with downloading and saving articles and comments — that is, readers’ ability to take delivery of what they have paid for.
This includes introducing barriers to commenters’ ability to save their own published comments and the rest of the commenter thread — the context in which they were made.
It used to be possible to save any article with its entire comment thread in a single sequence of clicks. That is now only a sad memory.
— Rearranging the order in which comments were actually posted in thread-saves, along with severely restricting the number of comments downloadable in a single cut-and-paste exercise. This is not merely frustrating but distorts the historical record, which matters to those of us — including scholars — who study for instance, the ebb and flow of sentiment on particular topics in the threads. Declarations of support for positions that oppose the paper’s politics can simply be eliminated from comment-rations.
— Stamping with an FT.com ownership (copyright) notice any saved comment contributed to the commons by readers writing to and for each other, a valuable public resource — even though we legally own the copyright to our own sentences and paragraphs.
— Discouraging commenters from getting to know each other.
In the first of the series of redesigns of FT.com witnessed by me since 2016 — including changes in the management of the site and policies affecting reader-commenters — subscribers stopped receiving detailed email notifications of replies to our comments as well as of any new comment in a thread if we had opted to be sent these. The e-notices had until then been delivered with copies of all the sentences to which a fellow-reader was responding; the whole text of any reply, as well as the screen name of the responder.
All these added immensely to the joys of commenting, especially for anyone like me far less interested in how many replies or recommendations my comments earn than in the screen identity of the recommender. In experiences on other sites, some responses were the start of exceptionally rewarding friendships. Being able to judge how active a thread was from new post alerts helped to decide timing for adding contributions to discussions or just larking around. The alerts often cheered up my email inbox on days when it was clogged with depressing subjects.
Since the policy shift, all we receive is a flag about a reply from a particular screen name. Nothing more. Another alteration introduced in those early years removed the ability to address a single comment to several screen names in a thread.
Why did the site managers do this? As they do not answer questions about their motives, you have to guess for yourself. Far from encouraging friendships and symphysis, the site operators are seemingly worried about the possibility of collective action by readers in any policy disagreement with management.
The overriding motive for switching to bare-bones email notices was obvious: to force commenters to return to the site as often as possible — to increase the click count used to attract advertisers, and to allow data collection. As almost everyone now knows, each twitch of our cursors, every placement of our clicks and even clues to our changing moods are being recorded and added to our profiles. Sparing us the need to return and log in again would mean giving up chances to fatten these data hoards.
— Blocking site visitors who pay to attend forums and other online events from using the privacy-preserving Safari browser to participate, and insisting that they use Chrome.
This happened to me after I had bought my ticket for one event. It took a time-wasting appeal to upper management to get my money back.
…
Because I put commenters at front and centre stage in my discussion brief for the Oxford Internet Institute in 2010, I have not been surprised to find comments being treated like precious metals — even if I couldn’t have guessed that this would have anything to do with data gathering for detailed profiling of citizens. Grist for the LLMs that are the mills of AI.
If further confirmation of the soaring value of the pearls of wisdom or pigeon-droppings in readers’ contributions were needed, it could be in a report last month that Chinese authorities are stigmatizing as mentally ill website visitors who read comments but seldom or never write any. Not altogether a shock, if true, when you consider that the Chinese have always had their own ways of going about compelling cooperation, and their AI mills are no less avid for everyone’s thoughts and feelings. It looks as if someone ordained that the psychological profiles of refuseniks must not be permitted to have holes.
…
I posted a comment on FT.com last summer about my fear that we are on our way to a hot, wet extinction by bain-marie beneath an article about AI. A reader using @Useparagraphsplease for a screen ID was delighted to remind me that in the famous experiment my metaphor came from, all the frogs were able to escape alive and unhurt.
But froggy dear, that took awareness and determined preventative action, which is my dearest hope for our species at the start of a new year.
Love,
p-G
…
Note: Readers of this post who feel that any criticism in it is unfair or inaccurate are invited to leave a comment (which will not be published until after its spambot check). Or, write to me directly at postgutenberg[at]gmail.com.
post-Gutenberg 