Nick Vincent

New post:

"Building a Data Pipeworks for Democratic AI: From Human Knowledge to Records to AI Systems"

In which I argue for zoomed out, human-centered model of data production (and circle back to arguing for old-school cybernetics?)

https://dataleverage.substack.com/p/building-a-data-pipeworks-for-democratic

Slowly getting back to blogging -- I have a somewhat lengthy post (of a somewhat unusual tone) that I think may be of interest to folks on here.

In short, this post is a sort of "Chapter 1" in a series of posts that serve as a rendition of the "modern AI is really reliant on / downstream of HCI" argument.

Or in the language of memes, "It's all HCI? Always has been".

Sharing the draft link here in case anyone wants to glance (or plod all the way through)

https://dataleverage.substack.com/p/25a36c8b-f537-4f05-9fe5-2c618b4d836b

Two related topics I think are critical to ongoing AI discussions: the beginning of a new era of negotiations between data creators and “data consumers” (artists, authors, newspapers, and more), and the potential enclosure of many pseudo-commons (a walled garden Internet).

So it was amazing to see coverage of these topics in two recent articles (and I was grateful to have a chance to comment)

On data leverage:
https://www.businessinsider.com/data-leverage-harry-potter-test-putting-value-data-ai-models-2023-8

On enclosure of the web:
https://www.businessinsider.com/ai-killing-web-grand-bargain-2023-8

@tarkowski@101010.pl @selenamarie@fiasco.social

I definitely think the notion of a Civic Option is a great idea. Certainly a lot of the work from the Wikimedia Research folks fits the bill here already (good recent overview here: https://research.wikimedia.org/report.html) -- I suppose there's a question here about the relative value of emphasizing LLM-specific avenues vs. enhancing existing recommendation and classifiers.

Certainly think giving orgs like WMF more AI capabilities seems good for society!

New post:
The WGA Strike is a Canary in the Coal Mine for AI Labor Concerns: Could Upcoming Data Legislation Enable a "Right to Data Strike"?

In a nutshell: Film, television, radio, and media writers are on strike, and their demands include restrictions on AI use. Studios have responded by threatening to use AI systems — trained using fruits of the labor of the very writers on strike — as AI Strikebreakers.

https://dataleverage.substack.com/p/the-wga-strike-is-a-canary-in-the

Microsoft is exploring Bing chat revenue sharing with content publishers: https://searchengineland.com/microsoft-explores-sharing-bing-chat-ad-revenue-with-publishers-395056

IMO - We're not far off from sharing LLM revenue with users too! For your consideration, a post on "Bing Rewards for the AI Age": https://nmvg.mataroa.blog/blog/bing-rewards-for-the-ai-age/

A while back I shared a "data leverage roundup": a file on GitHub providing links to support "collective action that involves withholding, deleting, manipulating, and/or transferring data".
It's now a website! https://www.datalevers.org

The plan: to collectively curate a set of tools for supporting -- or stopping -- the flow of data. Later on: also summarize empirical research and have an FAQ.

Current version is in construction, and really just a starting point. Seeking any and all feedback!

Double trouble warning: Birdsite crosspost + ChatGPT speculation

https://twitter.com/nickmvincent/status/1623405361188470784

Very interesting behavior found with reddit usernames of users who collaboratively count as high as possible via individual comments. The best example is: ChatGPT think the token " SolidGoldMagikarp" is the word "distribute". SolidGoldMagikarp is a now deleted counting hall-of-famer!

Curious what others think about this, I think it might be a big deal... one of the best "trap streets" for generative AI

New blog post: AI Technologies are System Maps, and You are a Cartographer: https://nmvg.mataroa.blog/blog/ai-technologies-are-system-maps-and-you-are-a-cartographer/

Still "beta testing" this idea, pushback welcome. BTW, you may be happy to hear that this post does *not* include the word "ChatGPT" (language models are mentioned, however...)

New op-ed with @hanlin@hci.social:

https://wired.com/story/chatgpt-generative-artificial-intelligence-regulation/

We lay out paths forward for public empowerment in the age of generative AI. Tell your friends: we’re all “expert chatGPT trainers”, and that gives us collective power!