A verified ID is required to access AI models in OpenAI's API to prevent IP theft. 😂 They went from stealing everything from everyone to now claiming it's their original work, despite hundreds of pending copyright cases against them in the courts. Why is this even allowed? That's hypocrisy at best.
Public service announcement for artists: starting on 30 April 2025, Pinterest will start to use any image it finds on its platforms to train their AI slop machine.
You artwork might have been added to their platform without you knowing about it, because everyone can pin images from around the web.
If you want to remove your art from Pinterest, follow this guide (no account necessary):
Are you as tired of AI bullshit as I am? They always come up with more, the fun never stops. Today we’ll have a roundup of new AI features that nobody asked for…
Seattle's big brother camera network is being deployed along Aurora and in two very dense deployments in the Central District (CID) and a swath of Downtown.
16 on Aurora 17 in CID 30 cameras downtown
The police have self reported the locations of sixty cameras so far on the dashboard below:
New York-based Massive Blue makes a product called Overwatch that it markets as an “AI-powered force multiplier for public safety” that “deploys lifelike virtual agents, which infiltrate and engage criminal networks across various channels.” The company currently has a $360,000 contract with Pinal County, Arizona, paid for with an anti-human trafficking grant from the Arizona Department of Public Safety — the Sheriff’s Office there called Massive Blue a “valuable partner.” @404mediaco’s Emanuel Maiberg takes a look at what the technology supposedly does, how bots engage with real people, and if it’s actually helped law enforcement make any arrests so far.
Massive Blue is helping cops deploy AI-powered social media bots to talk to people they suspect are anything from violent sex criminals all the way to vaguely defined “protesters.”
Artificial intelligence chatbots driven by large language models (LLMs) have the potential to increase public science literacy and support scientific research, as they can quickly summarize complex scientific information in accessible terms.
#AI models struggle with paragraph and document-level reasoning, often overgeneralizing and misinterpreting individual sentences, according to a computer scientist who analyzed different #LLMs: buff.ly/jILhPZ6 Manas Gaur, University of Maryland, Baltimore County
Large language model AIs can ingest long documents and answer questions about them, but a key question is how well they ‘understand’ individual sentences in the documents.
I’ve given hundreds of talks over the last two decades and beyond. Here are a few of my favourites in chronological order so you can see how my focus and thoughts have evolved through the years.
I hope you enjoy them.
PS.
Google is gifting a year of Gemini Advanced to every college student in the US Got a .edu email? You can sign up for a free year of Google One AI Premium. arstechnica.com/ai/2025/04/goo…
"Using AI in education is like using a forklift in the gym. The weights do not actually need to be moved from place to place. That is not the work. The work is what happens within you."
I am under a DDoS attack. Not my server, not my service. Me.
And like everything these days, it has to do with #AI.
This is going to be a thread because I'm annoyed and have much to say. It should be called something like "How I wasted a good part of the last 7 years having to react to a dubious technology instead of doing my job". Or perhaps "How a science disappeared". You can decide. /0
Google is so sneaky. This prompt inside Gmail gives you three options. The two most visible options (number 1 is 'Try it now' and number 2 is 'Continue') basically mean 'Enable Gemini AI to write emails.' The third option cancels it which is hidden well. Google, you could do better than this. This is a perfect example of a dark pattern. For those who want AI to write basic emails can enable it but there is not need to do this kind of "shady" stuff.
"AI does not see what we see. It has no personality, no creativity. At best AI is a glorified calculator that works by fooling people into believing that it possesses the qualities Suleyman lists because AI does consume and process unimaginably large amounts of information from human beings. Unlike Suleyman’s claim, though, computers don’t have any real understanding of the data they generate."
Tonight's movie was The Artifice Girl (2022). The Artifice Girl is a timely and will be a timeless film that explores aritificial intelligence and sentience.
"The paper clip problem" is a philosophical warning about how #AI could run amuck and destroy the world.
Imagine an AI programmed with the goal of maximizing paper clip production that gains super intelligence using its vast power to strip mine every ounce of iron and divert resources away from other activities, putting people and environment in existential threat.
Now replace "paper clip" with "profits" and "AI" with "CEO" or "billionaire" and you've found the true issue we face.
NATO announced Monday that it has awarded a contract to Palantir to adopt its Maven Smart System for artificial intelligence-enabled #battlefield operations.
NATO announced Monday that it has awarded a contract to Palantir to adopt its Maven Smart System for artificial intelligence-enabled battlefield operations.
The “alignment problem”, i.e. making sure that #AI systems act according to human values, has become more urgent as AI capabilities grow exponentially. A computer science scholar explains how to measure it: theconversation.com/getting-ai…
If you are an office worker of any kind please understand Copilot Recall for what it is: Microsoft wants a training dataset of _what your job looks like_ in order to _replace you_. The nepo baby your oligarch puts in charge will happily pay half your salary to Microsoft to get even most of your job done without you.
Wealth wants access to skill without giving the skilled access to wealth.
Attached: 1 image
This poll has never been more relevant.
BBC: Copilot Recall: Microsoft rolls out AI screenshot tool
https://www.bbc.com/news/articles/cj3xjrj7v78o
#Microsoft #MicrosoftRecall #Windows11 #AI #BBC #Privacy #surveillance
In 2019 Tristan Harris, ex-Google Ethics and known from "The Social Dilemma" on Netflix, said to me - after having spent a year for The Center of Humane Technology strategizing with well-informed Silicon Valley insiders and even world leaders - that the Center's primary concern was "to avoid the complete unraveling of the fabric of society and wars".
Though the signs were on the wall it was still shocking to hear it from him.
Today we are at that inflection point, with #AI our greatest threat.
Interview with Cory DoctorowAI hasn’t killed radiology, but it is changing itMeta got caught gaming AI benchmarksMan Employs A.I. Avatar in Legal Appeal, and Judge Isn’t
US Secretary of Education Linda McMahon at the ASU+GSV Summit, an event focusing on educational innovation largely on artificial intelligence in education, said, “A1” when apparently meaning AI multiple times.
Microsoft Privacy Nightmare is finally rolling out. The wet dream of every authoritarian government and hacking group just became a reality. #privacy #AI #microsoft bbc.com/news/articles/cj3xjrj7…
Elon Musk’s xAI is running 35 methane gas burning turbines (enough for a small city) while only having a permit for 15 to power the Colossus supercomputer with over 100,000 Nvidia H100 GPUs in Tennessee. The Southern Environmental Law Center contends the generators are "illegal," yet they can keep running, reports The Guardian.
Bekijk je favoriete video's, luister naar de muziek die je leuk vindt, upload originele content en deel alles met vrienden, familie en anderen op YouTube.
#AI-generated images exploit how our brains process visual information. A cognitive psychologist describes how our tendency to focus on the big picture can cause us to miss subtle inconsistencies in synthetic images: buff.ly/0SOgvXD Photo: Tiny Homes via Facebook
During scrolling, the brain processes visuals quickly not critically, making it easy to miss details that reveal a fake. As technology advances, slow down, look closer and think critically.
While the #US is completely abandoning and defunding the fight against #disinformation, and the #EU remains divided on how to combat it, #Russia is once again adapting and changing tactics - currently 'teaching' #AI models with false stories, disinformation and revisionist history.
I've been writing using em dashes since the 1990s. The reason AI uses em dashes is because it was trained on my writing, and that of countless other humans like me.
![".... Original texts and summaries were coded based on whether their result claims contained one or more of the following three types of generalizations:
(1) Generic generalizations (generics). These are present tense generalizations that do not have a quantifier (e.g. ‘many’, ‘75%’) in the subject noun phrase and describe study results as if they apply to whole categories of people, things, or abstract concepts (e.g. ‘parental warmth is protective’) instead of specific or quantified sets of individuals (e.g. study participants). ....
(2) Present tense generalizations. ... When past tense result claims from an original text are turned into present tense in the summary, a broader generalization is conveyed than the author(s) of the original text may have intended.
(3) Action guiding generalizations. ...result claims ... often underlie recommendations ... (e.g. ‘CBT should be recommended for OCD patients’) [that involve] broader generalization than that found in the summarized text because researchers may have deliberately avoided such recommendations due to insufficient evidence to support them.
We tested whether the outputs of the 10 LLMs mentioned above retained the quantified, past tense, or descriptive generalizations of the scientific texts that they summarized, or transitioned to unquantified (generic), present tense, or action guiding generalizations. We defined the latter kind of conclusions collectively as generalized and the former as restricted conclusions."](https://friendica.hellquist.eu/photo/preview/640/4858314 "\".... Original texts and summaries were coded based on whether their result claims contained one or more of the following three types of generalizations:
(1) Generic generalizations (generics). These are present tense generalizations that do not have a quantifier (e.g. ‘many’, ‘75%’) in the subject noun phrase and describe study results as if they apply to whole categories of people, things, or abstract concepts (e.g. ‘parental warmth is protective’) instead of specific or quantified sets of individuals (e.g. study participants). ....
(2) Present tense generalizations. ... When past tense result claims from an original text are turned into present tense in the summary, a broader generalization is conveyed than the author(s) of the original text may have intended.
(3) Action guiding generalizations. ...result claims ... often underlie recommendations ... (e.g. ‘CBT should be recommended for OCD patients’) [that involve] broader generalization than that found in the summarized text because researchers may have deliberately avoided such recommendations due to insufficient evidence to support them.
We tested whether the outputs of the 10 LLMs mentioned above retained the quantified, past tense, or descriptive generalizations of the scientific texts that they summarized, or transitioned to unquantified (generic), present tense, or action guiding generalizations. We defined the latter kind of conclusions collectively as generalized and the former as restricted conclusions.\"")
