Automatisez votre routine d'actualités quotidienne

Navigation

AccueilComment ça marcheContact

Légal

Politique de confidentialitéConditions d'utilisationMentions Légales
Made with ❤️ in France
Langue:

Revue de presse générée automatiquement avec

Nexarion
Voir tous les profils

Profil généré automatiquement

Nexarion

Un individu à l'aise avec la technologie, fasciné par l'intersection de l'intelligence artificielle, de l'IA générative et de la transformation numérique, à la recherche d'informations sur les dernières tendances technologiques et leur impact sur les industries. Il suit les progrès de l'IA et de ses applications, ainsi que les tendances technologiques émergentes.
Générative ai (30%)Artificial Intelligence (30%)Technology Trends (20%)Digital Transformation (20%)

Vous souhaitez recevoir chaque jour la revue de presse de ce profil ?

Gemini 3 Flash, AI governance, generative AI controversy, AI hardware...

Jeudi 18 décembre 2025 à 11:10

Generative AI & Content

Larian Studios’ GenAI backlash

The announcement that Larian Studios used generative AI for concept art and placeholder text in its upcoming Divinity title sparked a firestorm among fans and industry peers, who warned that such practices could erode creative jobs. The studio’s CEO, Swen Vincke, defended the limited use of AI while pledging that no AI‑generated assets would ship in the final game, a stance echoed by critics who called for clearer hiring standards. The episode underscores the growing tension between rapid AI adoption and traditional content creation pipelines. Wccftech Wccftech

AI Governance & Risk

Anthropic’s automated account shutdowns raise governance alarms

A Swiss cybersecurity firm’s experience with Anthropic—where an automated system abruptly cancelled its account and erased 80 % of its projects—highlights the perils of opaque AI‑driven enforcement. The incident, which required a paid reinstatement, prompted analysts to warn that enterprises must negotiate bounded disclosure clauses and maintain a human‑in‑the‑loop capability to avoid operational disruption. As Greyhound Research notes, nearly half of CIOs lack a response plan for such provider‑initiated shutdowns, turning a technical glitch into a governance crisis. Computer World

Instacart’s AI pricing tool draws FTC scrutiny

The Federal Trade Commission has issued a civil investigative demand to Instacart over its AI‑powered dynamic pricing engine, alleging potential violations of antitrust and consumer‑protection laws. Reuters‑cited sources say the probe focuses on whether the algorithm unfairly manipulates grocery prices, a concern amplified by growing regulatory attention to opaque AI decision‑making in commerce. The case may set a precedent for how AI‑driven pricing strategies are monitored across the retail sector. CNBC TechCrunch

AI Model Advances

Google rolls out Gemini 3 Flash as its new default model

Google announced that Gemini 3 Flash—a faster, more efficient successor to its previous models—will become the default engine in the Gemini app and power search results. Benchmarks show the model triples its score on the “Humanity’s Last Exam” test, closing the gap with the larger Gemini 3 Pro while delivering lower latency and reduced compute costs. The move signals Google’s intent to democratize high‑performance generative AI for both developers and end‑users. Ars Technica TechCrunch

AI in Software Development

AI‑generated code introduces more defects than human‑written code

A new analysis by CodeRabbit of 470 open‑source pull requests reveals that code produced by AI assistants exhibits significantly higher defect rates across logic, security, and performance categories. Complementing this, Tech Radar reports that AI‑generated snippets often contain hallucinations, forcing developers to spend additional time on manual reviews. The findings suggest that while AI accelerates coding speed, it also amplifies quality‑control challenges. Tech Radar DevOps.com

AI Hardware Race

AMD unveils Ryzen AI 7 445 with Gorgon Point architecture

AMD’s latest Ryzen AI 7 445 chip, spotted on Geekbench, features a novel 2 + 4 core configuration and integrates a Radeon 840M iGPU, positioning it as a cost‑effective option for AI workloads on the edge. The “Gorgon Point” architecture promises improved power efficiency and on‑device inference capabilities, reflecting the broader industry push to embed AI acceleration directly into consumer‑grade CPUs. Wccftech

Corporate Moves in AI

Amazon reshuffles its AGI leadership

Amazon confirmed that Rohit Prasad, head of its artificial general intelligence unit, will depart by year‑end, with senior AWS executive Peter DeSantis assuming control of a newly created division focused on advanced AI models, custom chips, and quantum computing. The transition underscores Amazon’s strategy to consolidate its AI ambitions under a single umbrella as it competes with rivals in the emerging AGI race. The Verge CNBC

OpenAI’s funding surge and talks of a Trainium partnership

OpenAI continues to attract heavyweight backers, securing a $1 billion investment from Disney and entering advanced talks with Amazon for a potential $10 billion infusion tied to the use of Amazon’s Trainium ASICs. Analysts warn that such deep pockets could widen the competitive gap between OpenAI and other AI labs, cementing its position at the forefront of large‑scale model development. The Information TechCrunch

AI‑Driven Devices

Wearables pivot from fitness to AI‑enabled experiences

The 2025 CES showcased a wave of AI‑infused wearables, from Meta’s Ray‑Ban glasses to audio‑only smart spectacles, positioning these devices as platforms for on‑device inference and personalized assistants rather than mere health trackers. Industry observers predict that the convergence of edge AI and form‑factor design will redefine how consumers interact with digital services on the go. The Verge

AI Environmental Impact

AI’s soaring water and electricity consumption raises sustainability concerns

A recent study estimates that AI workloads in 2025 consumed as much electricity as the global bottled‑water market and emitted carbon comparable to New York City, highlighting the hidden environmental cost of training and inference at scale. Researchers call for greater transparency from cloud providers and more energy‑efficient model architectures to mitigate AI’s growing carbon and water footprints. The Verge

Aller aux sources

16 sources citées

‘Larian Said They Were Doing Something Absolutely Everyone Is Doing and Got an Insanely Crazy Sh*tstorm’ – KCD 2 Dev Talks ‘AI Hysteria’

Wccftech

Game Writers Call Out Larian Studios For Poor Hiring Practices Amidst GenAI Backlash

Wccftech

Using AI to automatically cancel customers? Not a smart move

Computer World

Instacart shares drop on report that FTC is probing company over AI pricing tool

CNBC

Instacart’s AI-driven pricing tool attracted attention — now the FTC has questions

TechCrunch

Google releases Gemini 3 Flash, promising improved intelligence and efficiency

Ars Technica

Google launches Gemini 3 Flash, makes it the default model in the Gemini app

TechCrunch

AI-generated code contains more bugs and errors than human output

Tech Radar

Analysis Surfaces Rising Wave of Software Defects Traced to AI Coding Tools

DevOps.com

AMD Ryzen AI 7 445 Spotted On Geekbench — A 6-Core/12-Thread Gorgon Point Chip 2+4 Core Configuration

Wccftech

Amazon has a new leader for its ‘AGI’ group as it plays catch-up on AI

The Verge

Amazon says AI chief Rohit Prasad is leaving, Peter DeSantis to lead 'AGI' group

CNBC

OpenAI's Funding Advantage Over Rivals Grows

The Information

Amazon reportedly in talks to invest $10B in OpenAI as circular deals stay popular

TechCrunch

In 2025, wearables made a hard pivot to AI

The Verge

AI’s water and electricity use soars in 2025

The Verge