@ananny Drafting and tweaking is a good point. There will likely be paper mills that produce AI drafts and keep changing them until their AI generated origin is no longer visible. What human origin actually means (and its value) compared to AI origin of a draft needs some more thinking. Could there be a comparison to artistic random writing practices such as rolling dice?

What most intrigues me about ChatGPT, etc isn’t nefarious actors weaponizing it, students cheating, human labour replaced, or even datasets being flooded with machine-generated language (tho this’ll all probably happen to an extent).

It’s that it drives *drafting* — used to create first attempts that are tweaked, revised, edited in ways that will feel like non-ML authorship.

But the starting point will have been different.

I wonder/worry about the value & feeling of human starting points.

Important note for educators about now: it looks like GPT-2 output detectors can detect #chatgpt output pretty well.
https://huggingface.co/openai-detector
Edit: Changed my mind - it ID'd an excerpt from my own book as AI-written. A false positive could ruin a student's career.

Can somebody explain the #programming #language excitement of #chatGPT? I have a very rusty understanding of 1990s languages, so keep it layman accessible, please. Thanks!

“Meta's business model declared illegal in the EU according to WSJ. Facebook, Instagram and WhatsApp can no longer run personalized ads without user consent” https://noyb.eu/en/noyb-win-personalized-ads-facebook-instagram-and-whatsapp-declared-illegal

@kevlin

Christine Lemmer-Webber (@cwebber) described ChatGPT as Mansplaining As A Service, and honestly I can’t think of a better description. A service that instantly generates vaguely plausible sounding yet totally fabricated and baseless lectures in an instant with unflagging confidence in its own correctness on any topic, without concern, regard or even awareness of the level of expertise of its audience.

Entering the US public domain in 2023: Fritz Lang's silent film science-fiction classic *Metropolis*.

More info behind window 6 of our advent-style countdown calendar for works entering the #publicdomain on Jan 1st: https://publicdomainreview.org/features/entering-the-public-domain/2023/ #PDin2023 #film #metropolis

Interested in Indigenous approaches to AI & machine learning across disciplines?

Registration for #Indigenous IDEAS (Innovation, Data Economies & Artificial Intelligence Summit) is open! Event: Fri, 03/31/23 at Arizona State University. https://sols.asu.edu/ideas

One of the questions I often get is: well all this AI Bill of Rights stuff is well and good, but how does it align with the EU AI Act?

Among the many things I learnt during my time in government (thank you @cole_sci@twitter.com) is that there are processes in place to discuss these. 1/n

Thanks everyone! She found a job at Coiled.

So let’s see if Mastodon can be used to find jobs. Please boost for visibility! 🙏

Irina Truong is one of the best senior #Python developers I know. Her work is always outstanding and polished. She also contributes to open source: https://github.com/j-bennet. You want her in your team, trust me.

Unfortunately she was part of the 13% recently laid off by #Elastic. This is your luck! She’s now open to work: https://www.linkedin.com/in/irinatruong

In case you are looking for the #EU #Telecom Ministers Council adopted negotiation mandate for the #AIAct: https://data.consilium.europa.eu/doc/document/ST-14954-2022-INIT/en/pdf Here is an overview: https://www.euractiv.com/section/digital/news/eu-countries-adopt-a-common-position-on-artificial-intelligence-rulebook/
In the European #Parliament, we are still negotiating with two lead committees, but are making progress. #ArtificialIntelligence

Any reading tips on alternative approaches to process tracing? @politicalscience @sociology

The problem with GPT/LLM becoming so capable of copying human speech patterns is that it further erodes the public's trust in all discourse. If citizens can't differentiate between artificially created and authentic human speech, then they decide to completely tune out of all of it

1/ @ilyaz proposed to rename Artificial Intelligence in order to be more in line with what the technology actually does and help demystify it. As an #AI researcher who was never really interested in creating ‘intelligence’, I am sympathetic to this idea. Nevertheless, the field’s history has close connections with the goal of creating ‘intelligent’ systems. In this 🧵 my view on the question: To what extent is AI research about creating ‘intelligence’? >> https://mastodon.social/@ilyaz/109414907206515753

📢 Joint statement: The EU #AIAct must protect people on the move

In its current form, the AI Act does not adequately address and prevent harms stemming from the use of AI in the #migration context.

Today, AlgorithmWatch together with 191 organizations and individuals are sending a decisive signal to EU decision-makers: #ProtectNotSurveil migrants and #banBorderTech which is unjust, discriminatory and incompatible with #humanrights!

➡️ https://algorithmwatch.org/en/eu-must-protect-people-on-the-move/

@qruyant Well, it is reasonably realistic conversation with a particular kind of person. Third year student maybe?

Claim: ChatGPT is indistinguishable from an over-confident student.

Silicon Valley has a deep epistemological commitment to the idea that truth and plausible sounding bullshit are functionally equivalent.

It's become completely canonicalized by "fake it till you make it" factories like YCombinator becoming our industry's king makers.

The generative AI is wildly impressive and amusing, but also the product from the heart of the bullshit-is-truth nexus.

Not generally into "look at this amazing thing we've discovered is hidden inside an AI" type articles, but this is super interesting - using ChatGPT to emulate another simpler computer, connected to the internet: https://www.engraved.blog/building-a-virtual-machine-inside/

@thomasrenkert I guess this tells something about how innovations work. The capabilities can be lurking there in the background for quite a while. Then eventually someone comes up with a use for them that blows people away.