no-Flux

Lawsuit: A chatbot hinted a kid should kill his parents over screen time limits

13 décembre 2024 /

Two examples of interactions users have had with chatbots from the company Character.AI.

“ »It is simply a terrible harm these defendants and others like them are causing and concealing as a matter of product design, distribution and programming, » the lawsuit states.The suit argues that the concerning interactions experienced by the plaintiffs’ children were not « hallucinations, » a term researchers use to refer to an AI chatbot’s tendency to make things up. « This was ongoing manipulation and abuse, active isolation and encouragement designed to and that did incite anger and violence. »According to the suit, the 17-year-old engaged in self-harm after being encouraged to do so by the bot, which the suit says « convinced him that his family did not love him. »”

Source : Lawsuit: A chatbot hinted a kid should kill his parents over screen time limits : NPR

Cruise’s robotaxi service will likely shut down as GM pulls its funding

11 décembre 2024 /

Cruise Car

“General Motors said it would no longer fund its Cruise robotaxi service as it seeks to focus its spending on autonomous vehicle development specifically for personally owned vehicles. Now Cruise employees will be combined with GM’s internal teams working on advanced driver assist systems, like Super Cruise, as well as its project to develop autonomous vehicles to sell to customers for personal use. Ultimately, the project became too expensive for GM to justify the huge amounts of money spent to prop it up. And the automaker found it increasingly difficult to convince its shareholders that the money-losing operation would eventually pay off. The robotaxi subsidiary lost a staggering $3.48 billion in 2023 and has been seen by some as an albatross for the automaker, sucking up cash and lacking a clear path to profits.”

Source : Cruise’s robotaxi service will likely shut down as GM pulls its funding – The Verge

Un expert judiciaire accusé d’utiliser une IA pour générer son rapport sur les deepfakes

10 décembre 2024 /

Affichage de l'image en full screen

“…dans son texte, Jeff Hancock explique aussi que « la difficulté à identifier les deepfakes vient de la technologie sophistiquée utilisée pour créer des reproductions homogènes et réalistes de l’apparence et de la voix d’une personne. ». Pour l’affirmer, il s’appuie sur une étude qui « a montré que même lorsque les individus sont informés de l’existence des deepfakes, ils peuvent avoir du mal à faire la distinction entre un contenu réel et un contenu manipulé. Cette difficulté est exacerbée sur les plateformes de médias sociaux, où les deepfakes peuvent se propager rapidement avant d’être identifiés et supprimés. ». Sauf que cette étude ne semble pas exister.”

Source : Un expert judiciaire accusé d’utiliser une IA pour générer son rapport sur les deepfakes – Next

Brussels probes Google and Meta secret ads deal to target teens

10 décembre 2024 /

A montage of a teenager using a phone with the logos of Google and Meta in the background

“During the ad campaign, first revealed in an FT investigation published in August, staff at Google were helping Meta bypass the search company’s policies on how online advertising can be directed to minors. While Google bans ad personalisation for teens, the campaign for Instagram on YouTube deliberately pushed messages to a group of users labelled as “unknown” in its system. Google employees told Meta that it had internal data that showed this group skewed towards under-18s and was a way of “hacking” the audience safeguards in their system.”

Source : Brussels probes Google and Meta secret ads deal to target teens

Meet Willow, our state-of-the-art quantum chip

10 décembre 2024 /

“Willow has state-of-the-art performance across a number of metrics, enabling two major achievements. The first is that Willow can reduce errors exponentially as we scale up using more qubits. This cracks a key challenge in quantum error correction that the field has pursued for almost 30 years. Second, Willow performed a standard benchmark computation in under five minutes that would take one of today’s fastest supercomputers 10 septillion (that is, 1025) years — a number that vastly exceeds the age of the Universe. The Willow chip is a major step on a journey that began over 10 years ago.”

Source : Meet Willow, our state-of-the-art quantum chip

Disclose déplore la (cyber)surveillance hors norme d’une journaliste par la DGSI

10 décembre 2024 /

Affichage de l'image en full screen

“Quinze mois après sa garde à vue de 39 heures et la perquisition de son domicile par la Direction générale de la sécurité intérieure (DGSI), la journaliste Ariane Lavrilleux de l’ONG Disclose est convoquée, vendredi 17 janvier 2025, en vue d’une possible mise en examen pour « appropriation et divulgation d’un secret de la défense nationale ».Elle avait contribué, avec trois autres journalistes, à révéler en 2021 qu’une opération militaire secrète menée par la France en Égypte, baptisée « opération Sirli », avait « conduit à l’exécution arbitraire de centaines de civils égyptiens, le tout sur fond de vente d’armes ». La journaliste encourt une peine de cinq ans de prison et 75 000 euros d’amende.”

Source : Disclose déplore la (cyber)surveillance hors norme d’une journaliste par la DGSI – Next

Introducing ChatGPT Pro

5 décembre 2024 /

To highlight the main strength of o1 pro mode (improved reliability), we use a stricter evaluation setting: a model is only considered to solve a question if it gets the answer right in four out of four attempts ("4/4 reliability"), not just one.

“As AI becomes more advanced, it will solve increasingly complex and critical problems. It also takes significantly more compute to power these capabilities. Today, we’re adding ChatGPT Pro, a $200 monthly plan that enables scaled access to the best of OpenAI’s models and tools. This plan includes unlimited access to our smartest model, OpenAI o1, as well as to o1-mini, GPT-4o, and Advanced Voice. It also includes o1 pro mode, a version of o1 that uses more compute to think harder and provide even better answers to the hardest problems. In the future, we expect to add more powerful, compute-intensive productivity features to this plan. ChatGPT Pro provides a way for researchers, engineers, and other individuals who use research-grade intelligence daily to accelerate their productivity and be at the cutting edge of advancements in AI.”

Source : Introducing ChatGPT Pro | OpenAI

GenCast predicts weather and the risks of extreme conditions with state-of-the-art accuracy

5 décembre 2024 /

Two graphs compare GenCast and ENS for extreme heat and winds, showing GenCast provides higher relative economic value at 1- and 5-day lead times.

“Today, in a paper published in Nature, we present GenCast, our new high resolution (0.25°) AI ensemble model. GenCast provides better forecasts of both day-to-day weather and extreme events than the top operational system, the European Centre for Medium-Range Weather Forecasts’ (ECMWF) ENS, up to 15 days in advance. We’ll be releasing our model’s code, weights, and forecasts, to support the wider weather forecasting community.”

Source : GenCast predicts weather and the risks of extreme conditions with state-of-the-art accuracy – Google DeepMind

« Sans onde » : les petits mensonges des fabricants qui jouent avec la peur des parents

4 décembre 2024 /

« Sans onde » : les petits mensonges des fabricants qui jouent avec la peur des parents

“Jouer sur la peur des ondes pour faire vendre ? Ce n’est pas nouveau et nombreuses sont les équipes marketing à naviguer en eaux troubles sur ce sujet. Les enfants sont une cible de choix pour certains, qui n’hésitent pas à mentir pour mettre en avant un argument fallacieux. En voici quelques exemples frappants.”

Source : « Sans onde » : les petits mensonges des fabricants qui jouent avec la peur des parents – Next

« Older posts

Archives

Catégories

© 2024 no-Flux

Theme by Anders Noren — Up ↑