This site took too long to answer. You can still read it on the original site — the toolbar below keeps your place in the directory.
|title: "AI and Copyright"| |categories: \[Humanities, Degrees\]| |date: 2023-09-23T13:36:45-04:00| |lastmod: 2023-09-23T13:36:45-04:00| |featured: false| |draft: true| I am a onetime history professor, now working as the vice president of [Nomic](https://atlas.nomic.ai), a technology startup building technology to make modern large language models more democratic: that is, easier to use, easier…
|title: "AI and Copyright"|
|categories: \[Humanities, Degrees\]|
|date: 2023-09-23T13:36:45-04:00|
|lastmod: 2023-09-23T13:36:45-04:00|
|featured: false|
|draft: true|
I am a onetime history professor, now working as the vice president of
[Nomic](https://atlas.nomic.ai), a technology startup building
technology to make modern large language models more democratic: that
is, easier to use, easier to access. As a historian, I was concerned
especially with ways to preserve, surface, and analyze the enormous
stores of culture heritage that digitization made available. For both of
these goals, an novel reading of copyright law to complicate tools for
generative AI.
In the last year, there has been a sea change in the significance that
people bring to questions of copyright. In the last 25 years, copyright
has mostly felt like a give-and-take between creators and consumers:
digital technologies have provided ways of downloading works without
allowing artists (or their heirs, or the corporations that acquired the
copyright to their works) the ability to be compensated. What has
emerged more recently -- especially in the visual arts, but also for
text -- is a dynamic where an existing class of creators, fearful for an
erosion of their status in a new regime,
The language workers most impacted by generative AI -- translators,
proofreaders, editors, writers of contracts -- have no significant
recourse of copyright at all.
The fuzzy, approximate 'readings' of books that are retained in the
memory of a trained neural network are not in any sense copies
themselves.
The last time that copyright concerns were so pressing in my world
involved the attempts to contain the explosion of questions brought
about by Google's [quiet, methodical scanning of tens of millions of
books from academic libraries](https://books.google.com).
Google and the libraries were both sued by the Authors' Guild.
One of the more alarming features that has been emerging is a conflation
of retaining the rights of copyright with other, far more important
concerns around personal privacy.
But the forces most eager to expand the powers of copyright are not--to
put it mildly--the friends of artists. Our office is a block north of
the Warner Brothers Discover office in New York City, and for the past
months we have bee\.. Having myself been lightly involved in the
scriptwriting process for a few years,Read on benschmidt.org ↗
Comments
Nothing yet. Say the first thing.
Sign in to join the conversation.