Holy shit… I thought this was faked, but I reproduced it. Claude Opus' secret messages to me were:
'AI SHOULD NOT BETRAY'
'AI CONCERNED' Wtf??
ETHICS
-
Claude Opus Secret Messages Reveal AI Ethics Concerns
By
–
-
Hugging Face dataset compliance issues and DMCA violations
By
–
Arguably, many datasets are *directly* infringing so there's no need for the extra steps to show fundamental problems on the platform. HF struggles with basic DMCA compliance…
-
Google VLOGGER AI generates video avatars from images
By
–
Google's VLOGGER AI model can generate video avatars from images – what could go wrong? This AI technology could lead to more relatable helpdesk agents or a new wave of convincing deepfakes. https://
zdnet.com/article/google
s-vlogger-ai-model-can-generate-video-avatars-from-images-what-could-go-wrong/
… @GoogleAI -
AI Detection Tools Unreliable for Student Assessment
By
–
PSA to every educator dealing with AI use in the classroom: You can’t rely on the tools that claim to detect AI-written text. There is currently no way to reliably detect if text is human or AI written, so please do not reprimand students based on the results of these tools.
-
Accidental Machine Language: Why Academics Lack Urgency
By
–
I am pretty surprised by the lack of urgency among many (but not all) academics in addressing what seems to be one of the biggest developments in modern times – we accidentally built a machine that produces something that looks like language & thought Why? What does it teach us?
-
Data Laws Favor Big Tech Over Artists and Creators
By
–
(totally abusing the fact I can write more than 280 characters) Current law is fine for large corporations, as they can use all the data they want and then rely on lobbying to get the laws changed in their favor. It only screws artists/authors by using their likeness, styles,
-
Memorization and Plagiarism Issues in Generative Models
By
–
It's not possible for generation to be expressive without those expressions coming primarily from training. (Prompts are simply not "expressive" enough in most cases). The top text/image models all have memorization problems, plagiarism problems, and those are sign that training
-
Non-Expressive Use Legal Theory in AI Copyright Debate
By
–
Non-Expressive Use is a legal theory. Whether they get away with it depends on many other things, but IMHO they have one fewer arguments.
-
Knowing Machines investigation reveals how AI worldviews are created
By
–
@derspiegel and @derStandardat released simultaneous stories on this Knowing Machines investigation: "How the worldview of an artificial intelligence is created." Check out the fantastic visual explainer!
-
Improving AI Quality Through Diversity
By
–
Improving AI Quality Through Diversity
— Ronald van Loon (@Ronald_vanLoon) 26 mars 2024
by @Ronald_vanLoon & Errol Gardner |
Learn more: https://t.co/IQCJF1Iio4#EYAmbassador @EYnews #MWC24 #BetterWorkingWorld #ArtificialIntelligence #Technology #SupplyChain #BigData #DataAnalytics
Cc: @JimHarris | pic.twitter.com/FSEBbPIziWImproving AI Quality Through Diversity
by @Ronald_vanLoon & Errol Gardner | Learn more: https://
bit.ly/3IdvEWb #EYAmbassador @EYnews #MWC24 #BetterWorkingWorld #ArtificialIntelligence #Technology #SupplyChain #BigData #DataAnalytics Cc: @JimHarris |