Étiquette : artificial intelligence (Page 2 of 24)

Bank of England Warns of Impending AI Disaster

13 octobre 2025 / noflux

“Yet according to recent estimates, generative AI now accounts for roughly 40 percent of the United States’ gross domestic product. In other words, if the AI spending boom falls apart, it could take down the entire economy with it. Fund manager and former Morgan Stanley investor Ruchir Sharma warned in a recent piece for the Financial Times that the US economy has turned into “one big bet on AI.” “AI companies have accounted for 80 per cent of the gains in US stocks so far in 2025,” he wrote. “That is helping to fund and drive US growth, as the AI-driven stock market draws in money from all over the world, and feeds a boom in consumer spending by the rich.””

Source : Bank of England Warns of Impending AI Disaster

Les entreprises plongées dans le brouillard juridique du recrutement avec les IA émotionnelles

12 octobre 2025 / noflux

“Entré en vigueur en juin 2024, le premier volet du règlement européen sur l’intelligence artificielle (AI Act), qui porte sur les IA interdites, intègre dans le cadre du travail ces outils de déduction des émotions. « La base scientifique des systèmes d’IA visant à identifier ou à déduire les émotions suscite de vives inquiétudes, d’autant plus que l’expression des émotions varie considérablement d’une culture à l’autre et d’une situation à l’autre, voire au sein d’un même individu, souligne le règlement. Parmi les principales lacunes de ces systèmes figurent leur fiabilité limitée, leur manque de spécificité et leur capacité limitée de généralisation. »
La Commission nationale de l’informatique et des libertés met d’ailleurs en garde les entreprises : « Ce n’est pas parce que ces solutions sont en vente qu’elles sont légales et conformes d’un point de vue de la protection des données », prévient Eric Delisle, chef de service juridique de cette entité, en rappelant que c’est aux entreprises qu’il revient de s’assurer sur ce point qu’elles sont bien dans la légalité. L’avertissement vaut d’autant plus que ce texte encore récent invite à la prudence dans son interprétation.”

Source : Les entreprises plongées dans le brouillard juridique du recrutement avec les IA émotionnelles

China tests underwater data centres to reduce AI carbon footprint

6 octobre 2025 / noflux

An underwater data centre being developed by Chinese maritime technology company Highlander is seen under construction at a shipyard in Nantong, in China's eastern Jiangsu province, on September 11, 2025

“The technology was trialled by Microsoft off the coast of Scotland in 2018, but the Chinese project, to be sunk in mid-October, is one of the world’s first commercial services of its kind.It will serve clients such as China Telecom and a state-owned AI computing company and is part of a broader government push to lower data centres’ carbon footprint.“Underwater facilities can save around 90 per cent of energy consumption for cooling,” said Yang, vice-president of Highlander.Projects like this are currently focused on showing “technological feasibility”, said expert Shaolei Ren from the University of California, Riverside.”

Source : China tests underwater data centres to reduce AI carbon footprint | South China Morning Post

Intelligence artificielle, données, calculs : le rapport final du Shift Project

1 octobre 2025 / noflux

Intelligence artificielle, données, calculs :
L’insoutenable croissance de l’offre et des usages

“En plein « phénomène IA », The Shift Project a choisi d’éclairer le sujet avec son prisme : la vision physique. Pour ce faire, nous avons étudié une composante clé des infrastructures du numérique, la filière centre de données, et la manière dont elle se construit en interaction avec l’intelligence artificielle, principal déterminant de ses dynamiques aujourd’hui.
Le programme numérique du Shift Project (voir l’ensemble de nos travaux) mène et documente depuis plusieurs années une réflexion sur les pratiques et actions qui permettent de limiter les impacts environnementaux directs et indirects du numérique, sans empêcher l’effet net des potentiels leviers qu’il propose en matière de transition écologique.
Le numérique est un secteur non négligeable : il représentait déjà 3 à 4 % des émissions mondiales en 2020 (The Shift Project, 2021), soit du même ordre que l’intégralité des poids lourds dans le monde (IEA, 2021) et avec une augmentation de 6 %/an en moyenne de cette empreinte. À l’échelle française, il représentait 4,4 % de l’empreinte carbone du pays en 2022 (ADEME, 2025).
Ce nouveau projet trace le contour de la manière dont le déploiement généralisé de l’IA infléchit ces dynamiques déjà insoutenables. Il éclaire les pistes à suivre pour réorienter ”

Source : Intelligence artificielle, données, calculs : le rapport final du Shift – The Shift Project

Neon, the No. 2 social app on the Apple App Store, pays users to record their phone calls and sells data to AI firms

25 septembre 2025 / noflux

“In a brief test by TechCrunch, Neon did not offer any indication that it was recording the user’s call, nor did it warn the call recipient. The app worked like any other voice-over-IP app, and the caller ID displayed the inbound phone number, as usual. (We’ll leave it to security researchers to attempt to verify the app’s other claims.)
Neon founder Alex Kiam didn’t return a request for comment.
Kiam, who is identified only as “Alex” on the company website, operates Neon from a New York apartment, a business filing shows.
A LinkedIn post indicates Kiam raised money from Upfront Ventures a few months ago for his startup, but the investor didn’t respond to an inquiry from TechCrunch as of the time of writing.”

Source : Neon, the No. 2 social app on the Apple App Store, pays users to record their phone calls and sells data to AI firms | TechCrunch

How People Use ChatGPT

17 septembre 2025 / noflux

“Despite the rapid adoption of LLM chatbots, little is known about how they are used. We document the growth of ChatGPT’s consumer product from its launch in November 2022 through July 2025, when it had been adopted by around 10% of the world’s adult population. Early adopters were disproportionately male but the gender gap has narrowed dramatically, and we find higher growth rates in lower-income countries. Using a privacy-preserving automated pipeline, we classify usage patterns within a representative sample of ChatGPT conversations. We find steady growth in work-related messages but even faster growth in non-work-related messages, which have grown from 53% to more than 70% of all usage. Work usage is more common for educated users in highly-paid professional occupations. We classify messages by conversation topic and find that “Practical Guidance,” “Seeking Information,” and “Writing” are the three most common topics and collectively account for nearly 80% of all conversations. Writing dominates work-related tasks, highlighting chatbots’ unique ability to generate digital outputs compared to traditional search engines. Computer programming and self-expression both represent relatively small shares of use. Overall, we find that ChatGPT provides economic value through decision support, which is especially important in knowledge-intensive jobs. ”

Source : How People Use ChatGPT | NBER

AI Sovereignty Is a Spectrum, Not a Switch

16 septembre 2025 / noflux

“To have full AI sovereignty, you’d need to control absolutely every element of the AI pipeline. Think about what that really means. You’d need to develop and build your own chips; not just design them, but actually manufacture them. You’d need to build and run your own models. Your data pipeline would need to be completely under your control. And so on.And all of this – from the rare earth elements in your hardware to the energy powering your data centers – would need to be independent from external forces. Maybe the United States can achieve this. China might get close. But for smaller countries? For Switzerland, Belgium, or really any nation that isn’t a global superpower? It’s just not realistic.”

Source : AI Sovereignty Is a Spectrum, Not a Switch

How Simplify in the Google app uses AI to make complex text easier to understand

9 septembre 2025 / noflux

A flowchart illustrates an iterative loop for improving text simplification. An original text is simplified and then evaluated on readability, completeness and entailment. Based on the evaluation, prompts are refined and ranked to continue the loop.

“The path to Simplify began in the specialized world of medicine, where no detail should be spared. « Doctors sometimes use language that’s purposefully obscured to reduce patient anxiety or preserve privacy, » says Diego Ardila, a Google Research software engineer. « They might say, ‘The patient is undergoing emesis,’ which just means they’re vomiting. Sometimes there’s a use for that inside a hospital, but other times it actually gets in the way.”The team built an internal simplification demo, and started testing it on text outside of medicine. “We found that it just kept working because the underlying AI models are general-purpose,” Diego says. “Seeing its potential for broader applications, we shared it with other teams.””

Source : How Simplify in the Google app uses AI to make complex text easier to understand

YouTube secretly used AI to edit people’s videos. The results could bend reality

27 août 2025 / noflux

Illustration by Serenity Strull/ Getty Images An illustration of a woman's face being altered and skewed by-the-pixel on a YouTube video (Credit: Illustration by Serenity Strull/ Getty Images)

“YouTube did not respond to the BBC’s questions about whether users will be given a choice about AI tweaking their videos.It’s certainly true that modern smartphones come with built-in AI features that can enhance image and video quality. But that’s an entirely different affair, according to Samuel Woolley, the Dietrich chair of disinformation studies at the University of Pittsburgh in the US. « You can make decisions about what you want your phone to do, and whether to turn on certain features. What we have here is a company manipulating content from leading users that is then being distributed to a public audience without the consent of the people who produce the videos. »”

Source : YouTube secretly used AI to edit people’s videos. The results could bend reality

Elon Musk’s xAI Published Hundreds Of Thousands Of Grok Chatbot Conversations

20 août 2025 / noflux

“Today, a Google search for Grok chats shows that the search engine has indexed more than 370,000 user conversations with the bot. The shared pages revealed conversations between Grok users and the LLM that range from simple business tasks like writing tweets to generating images of a fictional terrorist attack in Kashmir and attempting to hack into a crypto wallet. Forbes reviewed conversations where users asked intimate questions about medicine and psychology; some even revealed the name, personal details and at least one password shared with the bot by a Grok user. Image files, spreadsheets and some text documents uploaded by users could also be accessed via the Grok shared page. ”

Source : Elon Musk’s xAI Published Hundreds Of Thousands Of Grok Chatbot Conversations