Big Tech's free use of copyrighted work to train AI devalues creators, crowns a new techno-elite

Time of India

25-04-2025

Of all the lawsuits that continue to snap at Meta's heels today, the most interesting one concerns how the tech giant used millions of
pirated books
to train its
LLAMA algorithms
.
#Pahalgam Terrorist Attack
India pulled the plug on IWT when Pakistanis are fighting over water
What makes this India-Pakistan standoff more dangerous than past ones
The problem of Pakistan couldn't have come at a worse time for D-St
Plaintiffs in 'Richard Kadrey et al v. Meta' have filed a motion, accusing the company of having used 'millions of books and other copyrighted works... for free and without consent from the rightsholders because it does not want to pay for them'. Even more interesting than the charge, however, is how Meta has sought to defend itself through sundry confidential internal exchanges that have been released to the public domain: it has accepted culpability but denied liability for copyright infringement, claiming that the 7 mn books used to train its LLM constituted '
fair use
' of already-compromised material.
Fair use of intellectual work is not new. In ancient India, the Vedas were considered shruti (that which is heard) and apauruseya (not of any man, impersonal), because it was more important to ensure the unbroken continuation of an oral - and later written - tradition, whose authorship was less important than its preservation. It was the same elsewhere, though sometimes works such as The Iliad and The Odyssey were loosely ascribed to a poet named Homer.
Play Video
Pause
Skip Backward
Skip Forward
Unmute
Current Time
0:00
/
Duration
0:00
Loaded
:
0%
0:00
Stream Type
LIVE
Seek to live, currently behind live
LIVE
Remaining Time
-
0:00
1x
Playback Rate
Chapters
Chapters
Descriptions
descriptions off
, selected
Captions
captions settings
, opens captions settings dialog
captions off
, selected
Audio Track
default
, selected
Picture-in-Picture
Fullscreen
This is a modal window.
Beginning of dialog window. Escape will cancel and close the window.
Text
Color
White
Black
Red
Green
Blue
Yellow
Magenta
Cyan
Opacity
Opaque
Semi-Transparent
Text Background
Color
Black
White
Red
Green
Blue
Yellow
Magenta
Cyan
Opacity
Opaque
Semi-Transparent
Transparent
Caption Area Background
Color
Black
White
Red
Green
Blue
Yellow
Magenta
Cyan
Opacity
Transparent
Semi-Transparent
Opaque
Font Size
50%
75%
100%
125%
150%
175%
200%
300%
400%
Text Edge Style
None
Raised
Depressed
Uniform
Drop shadow
Font Family
Proportional Sans-Serif
Monospace Sans-Serif
Proportional Serif
Monospace Serif
Casual
Script
Small Caps
Reset
restore all settings to the default values
Done
Close Modal Dialog
End of dialog window.
by Taboola
by Taboola
Sponsored Links
Sponsored Links
Promoted Links
Promoted Links
You May Like
Villas For Sale in Dubai Might Surprise You
Villas In Dubai | Search Ads
View Deals
This continued until only a few hundred years ago, when authorship and intellectual ownership became things of worth. Not the least of these reasons was the need to establish authenticity, ownership and factual origin. Not the most significant (or perhaps the most significant) was the fact that intellectual output attracted a monetary value, irrevocably dwarfing that associated with manual and martial labour.
Since that time, works of intelligence and creativity have only gained currency. Today, when success often hinges on ideas, innovation, entrepreneurship and opinions, the protection of intellectual copyright is more critical than ever before.
Live Events
Two questions, therefore, arise:
Why does Meta want to upend the status quo?
How does Meta believe it can get away with this?
The first question is easy enough to answer. It has declared that LLAMA is 'highly transformative... and the use of copyrighted materials is vital to the development of the company's
open-source AI models
'. And, yet, the
Association of American Publishers
claims that 'the systematic copying of textual works, word by word, into an LLM' - without, among others, critical commentary, search functionality and digital interoperability - cannot be considered 'transformative under fair use precedents'.
The veracity of this last statement is easily established by examining the way in which Meta seeks to commodify these works. Meta is not interested in making its AI tool a ready-reference library for the sake of access and preservation. Instead, as recently uncovered written communication between researchers has shown, the use of pirated texts, especially works of modern fiction, was 'easy to parse' for LLM training.
However, where things could take a bizarre turn, according to several lawyers and former employees, is when original works are used to produce new output IOD (instantly on demand), including unlicensed sequels, derivative literature, wholly fallacious background material, and even entirely new work in the style of other published work - all of which would result in transforming creative output into a cheap asset, trivialising individual authorship, and making authentic and original intellectual pursuit superfluous in the long run.
Which begs an answer to the second question, to which Meta has responded: by not paying for it. After initially investigating the possibility of entering into licence fee agreements with authors and publishers, it abandoned this endeavour because of cost, time and resource considerations.
Subsequently, it has not only sought refuge in diverting liability for copyright infringement to those involved in book piracy but has also invoked the power of mathematics - reluctant arbiter of truth - to show that an individual work, however large or illustrious, could never enhance an LLM's performance by more than 0.06%, 'a meaningless change no different from noise'.
Thus, Meta sees no reason to pay individuals, since they have little of value to exchange with the company - a superb piece of casuistry that would have cheated Shakespeare out of royalties accruing from his 37 plays and 154 sonnets, because they are statistically insignificant.
Ultimately, this may be a case of history defending its right to repeat itself.
At some point, the Vedic system of open-source education based on ability gave way to a more stratified hierarchy, where merit was sacrificed to birthright because Brahmins wished to retain the prestige and wealth associated with acquired knowledge for themselves. As the caste system took hold, the jealously guarded ritual was elevated at the expense of more equitable learning.
Today, Meta and others - OpenAI has even asked Donald Trump to allow
intellectual theft
to stop China from stealing a march on America - seek to use copyrighted material without compensation to build future go-to knowledge resources, from which they hope to capture unimaginable wealth and power. Authentic and creative work will, thus, be made redundant and worthless, to credit and aggrandise a new techno-hierarchy.
Let us pray that the courts prevail, and history fails in its defence.