Miguel Afonso Caetano · @remixtures
599 followers · 2392 posts · Server tldr.nettime.org

: "An emerging type of AI system is a ‘foundation model’, sometimes called a ‘general-purpose AI’ or ‘GPAI’ system. These are capable of a range of general tasks (such as text synthesis, image manipulation and audio generation). Notable example are OpenAI’s GPT-3 and GPT-4, foundation models that underpin the conversational chat agent ChatGPT.

Because foundation models can be built ‘on top of’ to develop different applications for many purposes, this makes them difficult – but important – to regulate. When foundation models act as a base for a range of applications, any errors or issues at the foundation-model level may impact any applications built on top of (or ‘fine-tuned’) from that foundation model.

As these technologies are capable of a wide range of general tasks, they differ from narrow AI systems (those that focus on a specific or limited task, for example, predictive text or image recognition) in two important respects: it can be harder to identify and foresee the ways they can benefit people and society, and it is also harder to predict when they can cause harm."

adalovelaceinstitute.org/resou

#ai #generativeAI #gpai #foundationmodel

Last updated 1 year ago

I still don't grok why systems are exciting.

I get the pursuit to understand human mind/intelligence. We can consider humans to be generally intelligent - we can learn and adapt to so many things.

But, for , I want to see improved support for various human tasks.

Current crop of are not human-like & don't serve as explanatory theories of the human mind.

Worse, their fame as 'generally intelligent' moves focus from specific usecases towards 'magic'.

#generapurposeai #gpai #AGI #ai #ml

Last updated 1 year ago

Miguel Afonso Caetano · @remixtures
373 followers · 1140 posts · Server tldr.nettime.org

: "Regulation should avoid endorsing narrow methods of evaluation and scrutiny for GPAI that could result in a superficial checkbox exercise. This is an active and hotly contested area of research and should be subject to wide consultation, including with civil society, researchers and other non-industry participants. Standardized documentation practice and other approaches to evaluate GPAI models, specifically generative AI models, across many kinds of harm are an active area of research. Regulation should avoid endorsing narrow methods of evaluation and scrutiny to prevent this from resulting in a superficial checkbox exercise."

ainowinstitute.org/news/gpai-i

#ai #gpai #generativeAI #eu #AIAct

Last updated 1 year ago

Rob Williamson · @eswag
116 followers · 1435 posts · Server dju.social

This is an interesting polemic about .

bigthink.com/starts-with-a-ban

It reminds me of expressions of "legitimate concern about ", or, as it is more concisely termed, .

I still think my career in is heading for an abrupt end at the neurons of a .

#concerntrolling #immigration #racism #software #gpai

Last updated 1 year ago

Alexander Schäfer ツ · @schaefer
32 followers · 385 posts · Server berlin.social

Bei wird darüber nachgedacht, wie |en bei der Erstellung von Beiträgen eingesetzt werden könnten. Gründer Jimmy Wales kann nicht einschätzen, wie weit Wikipedia davon noch entfernt ist. Größtes Problem: die gelegentlich fehlende |treue.

t3n.de/news/wikipedia-kuenftig

#wikipedia #ki #textgenerator #fakten #chatbot #llm #gpai #chatgpt

Last updated 1 year ago

Alexander Schäfer ツ · @schaefer
29 followers · 422 posts · Server berlin.social
Miguel Afonso Caetano · @remixtures
301 followers · 693 posts · Server tldr.nettime.org

: "The EU lawmakers spearheading the work on the AI Act pitched significant obligations for providers of large language models like ChatGPT and Stable Diffusion while seeking to clarify the responsibilities alongside the AI value chain.

The AI Act is a flagship EU legislation to regulate Artificial Intelligence based on its capacity to cause harm. A big question mark in the negotiations of the legislative proposal is how to deal with General Purpose AI (GPAI), large language models that can be adapted for various tasks.

The offices of the European Parliament’s co-rapporteurs, Dragoș Tudorache and Brando Benifei, shared on Tuesday (14 March) a first draft on this sensitive topic, proposing some obligations for the providers of this type of AI models and responsibilities for the different economic actors involved."

euractiv.com/section/artificia

#eu #ai #generativeAI #gpai #chatgpt

Last updated 2 years ago

Ian Brown :fedi: · @1br0wn
1920 followers · 1579 posts · Server eupolicy.social

Loving Algorithmic Black Swans by Kolt (forthcoming). “The world’s regulatory “superpower” and first-mover in AI regulation is failing to establish appropriate safeguards around the most consequential AI technology.” (p.34) papers.ssrn.com/sol3/papers.cf

#AIAct #gpai #brusselseffect

Last updated 2 years ago

Lupposofi · @lupposofi
87 followers · 1557 posts · Server mastodontti.fi

Nostan tuon 33-sivuisen raportin erikseen esille: corporateeurope.org/sites/defa. Siitä selviää, että "AI is an inherently political tool that affects the distribution of power in society" (s. 7), GPAI (kuten ChatGPT) on jäänyt eettisen ja muunkin harkinnan katveeseen, kiitos em. vinon (a)siantuntijaryhmän ja myöhemmän lobbauksen (mainitaan nimeltä mm. Aura Salla), eikä salainen trilogue(?) rohjenne suosittaa sellaista korkean riskin kategoriaan.

#trilogue #eu #ethics #chatgpt #gpai

Last updated 2 years ago