-
Vibe hacking: when users manipulate the behavior of generative AI
Since the emergence of large language models, researchers and developers have focused their attention on securing their outputs. Filtering, alignment, and control mechanisms are designed to make AI “useful, honest, and harmless.” However, a new form of circumvention has emerged—one that is more subtle than jailbreaks…
-
"Temporary Chats" on Gemini: A Step Toward More Privacy-Friendly AI
The emergence of generative artificial intelligence, capable of interacting in natural language, is profoundly transforming our relationship with technology. These conversational tools are becoming everyday assistants, whether for work, learning, creating, or staying informed.
-
ChatGPT is stepping up a gear: OpenAI unveils GPT-5
The announcement had been anticipated for over a year. On August 20, 2025, OpenAI officially unveiled GPT-5, the latest version of its flagship language model, which is integrated into ChatGPT and the OpenAI API. This release comes amid intense competition, with Anthropic’s Claude 3.5, Gemini 2 from…
-
dAIsy: Defensive Artificial Intelligence to Thwart Phone Scammers
The voice is soft, slightly hesitant. She talks about her family, asks for details, deliberately loses her train of thought… but she isn’t a real elderly person. On the other end of the line, a scammer thinks he’s tricked a vulnerable target. He has no idea that he’s wasting his time…
-
Bringing Color to the Past: How Google’s Artificial Intelligence Enhances Black-and-White Photos
Old black-and-white photos have a special charm: they freeze the past, capture an era, and evoke memories that are often personal or family-related. But they can sometimes be difficult for younger generations to appreciate, as they are unfamiliar with their visual context. Thanks to recent advances in artificial intelligence, it…
-
Gemini in the Apple ecosystem: toward a smarter Siri, made by Google
Apple and Google have long embodied an iconic rivalry in the tech world. One advocates for vertical control over the user experience, while the other favors a more open approach focused on online services. But in June 2025, at its WWDC conference, Apple surprised everyone by announcing the integration of…
-
Google Search's AI mode is expanding to 180 countries: a new era in search
Google Search and Generative AI: A Historic Turning Point With more than 8.5 billion searches per day, Google Search remains the primary gateway to online information for the majority of internet users worldwide. But since 2023, the search engine’s dominance has been challenged by virtual assistants…
-
Meta and Midjourney: A Strategic Partnership to Revolutionize AI-Powered Images and Video
In this context, the partnership announced in the summer of 2025 between Meta (the parent company of Instagram, Facebook, WhatsApp, and Threads) and Midjourney marks a strategic turning point. It is no longer simply a matter of generating individual images, but of building a comprehensive multimodal infrastructure capable of transforming a text query into an image, video, or…
-
Perplexity AI offers a revenue-sharing model with media outlets: toward a new model of collaboration
The rise of generative artificial intelligence has profoundly transformed the way information is produced, disseminated, and accessed. Tools such as ChatGPT, Gemini, and Perplexity allow users to obtain synthetic responses to complex queries, often drawing on journalistic or encyclopedic content.
-
GPT-OSS: OpenAI Releases Its First Open-Source Models Since 2019
Since 2019, OpenAI had moved away from its initial commitment to open source, opting instead for a more commercial and controlled strategy centered on its proprietary models (GPT-3, GPT-4, and then GPT-4o). It was therefore with some surprise that the AI community welcomed, in July 2025, the launch of…
-
MLE-STAR: Google’s approach to effectively structuring machine learning engineering
Drawing on its extensive experience in deploying AI at scale, Google offers a methodological solution to this challenge with the MLE-STAR framework. Designed as a synthesis of best practices in software engineering tailored to machine learning, this framework aims to structure AI projects in a more reliable and modular way…
-
Qwen3: Alibaba’s model that challenges OpenAI and DeepSeek in mathematics and coding
While large language models are dominated by the United States, China is gradually strengthening its position in the field of advanced artificial intelligence. With Qwen3, Alibaba aims to offer a competitive model in the strategic areas of mathematical reasoning and code generation.
-
ChatGPT Agent: OpenAI Introduces an AI Capable of Planning, Executing… and Learning
Since the emergence of large language models, conversational artificial intelligence has made rapid strides in understanding, content generation, and interactivity. But until now, assistants like ChatGPT, Claude, and Gemini have been limited to the role of a conversational partner, unable to take direct action within a digital environment.
-
AI-based age verification in tobacco shops: The CNIL strongly opposes it
In recent years, artificial intelligence systems capable of estimating a person’s age based on their face have become increasingly common in retail settings. Using so-called “smart” cameras, these devices analyze facial features in real time to estimate an age range, without necessarily identifying the individual.
-
AI and Speech: Voxtral, Mistral’s Open-Source Response to Large Language Models
Artificial intelligence is no longer limited to vision or text. In recent years, speech has become a strategic area of research, where technical, commercial, and political issues intersect. While automatic transcription has made significant strides, the ability of machines to truly understand language…