Menu
Ireland's political and cultural magazine
56 April-May 2025
April-May 2025 PB
Rise And Fall of
the Machine
For media, trustworthy news trumps
vogues for AI, Crypo, Quantum,
click-bait opinions and the rest
By Gerard Cunningham
E
very year the Nieman Lab, a part of
Harvard University funded by an
American philanthropic group, asks
a wide range of editors, publishers
and journalists for their predictions
for the coming year.
The 2025 predictions reveal a mix of
anxieties, firstly about the continued failures
in reporting critical issues such as climate
chance, and about the imminent return to
oce of a vengeful Donald Trump, but also
about the ongoing impact of so-called
“Artificial Intelligence” projects and Large
Language Models (LLM).
By February, some of these predictions were
playing out, as shown when publishers in the
UK from the Guardian to the Daily Mail found
themselves on the same side of an issue for
once, united in opposition to copyright grabs
from the AI promoters hungry for content to
feed into their learning models.
As a freelance writer, I have more than once
been told on seeing a story I wrote lifted by a
rival publication that “you cannot copyright the
news”, so forgive me if I am not too pushed
about the furious newsroom reactions to AI
copyright grabs.
While the impact on individual creatives is
undeniable, and smaller independents now
have to add LLMs and image generators to the
threats to their business, in addition to
unscrupulous lifting of images and words to
create unlicensed merchandise, LLMs are
Newsrooms look at the billions spent on
these projects and despair, as most of them
are caught in an endless spiral of cutbacks
despite already having systems to publish
information using natural intelligence, if
readers were willing to pay for it
mostly a concern for corporations.
As corporations agree terms with data
scrapers, dropping threatened lawsuits in
return for agreements paying royalties for
access to their archives, precious little of that
seems to trickle down to the writers,
photographers and others creating the words
and images being hoovered up.
There are signs that the AI hype train is
running out of steam, so newsrooms may be
wise to take in as much cash as they can while
it’s available.
As outlined at length by Ed Zitron in his
newsletter, (most notably in The Subprime AI
Crisis, published in September 2024) the
industry is a Potemkin village built on sand,
burning energy and losing money with every
query. After all the hype and promise, LLMs
have no useful applications, and are widely
regarded with contempt by consumers, who
are forced to endure unasked “upgrades” to
add AI to products, for which they are then
charged extra.
Not only have the Silicon Valley companies
been unable to find useful applications for
Clippy on Steroids, but their existing products
have been compromised. Hardly a day goes by
without another example of a basic mistake
from Google, reciting garbled nonsense in
response to a question, whether its lethal food
recipes or gibberish science news.
What happens next is up for debate. Before
AI/LLM, the last great grift was cryptocurrency.
The Trump administration has announced
plans for a “US Crypto Reserve”, though
whether it materialises remains to be seen.
With several “whales” left holding large
amounts of cryptocurrency they cannot cash
out without crashing the remaining markets,
this amounts to what American political
commentator David Frum described as the US
Treasury acting as “the greatest fool of last
resor t”.
Whether Trump and Musk could arrange a
similar bailout for AI backers is hard to say,
though it’s not impossible to imagine directives
to the Pentagon forcing it to spend billions
buying “AI analysis tools” or suchlike.
Meanwhile, the best guess from most
observers is that the next big thing from the
Silicon Valley scammers will be Quantum
Computing. More billions, poured into yet more
companies promising miraculous vapourware,
all just two years away if you invest now, only
a couple of billion more dollars to buy more
expensive and power hungry servers, and pay
no attention to the megatonnes of fossil fuels
burned to power them and pump out hot air
and greenhouse gases.
Newsrooms must look at the billions spent
on these projects and despair, as most of them
are caught in an endless spiral of cutbacks, lay-
offs, circulation declines and shrinking
pagecounts, despite already having in place
established systems to uncover, publish and
distribute verifiable information using natural
intelligence, if only readers were willing to pay
for it.
Yet paradoxically, the news sections which
attract most attention are those most removed
from the fact-gathering and presentation
process. Opinion pages, often populated by a
class of professional agitators trolling for
attention, routinely generate more heat than
light. Desperate for the pageviews that edgy
opinion pieces bring, newsrooms burn through
their most valuable asset, the credibility and
trust built up over decades, and wonder why
readers continue to cancel subscriptions.
MEDIA
