Modelos, agentes, investigación y negocio de la IA.
OpenAI has identified an influence operation attributed to China that used its own chatbot to create social media posts. The campaign sought to fuel opposition to AI data centers and criticize Trump's tariffs, though it had little real impact.
Coinbase has introduced Coinbase for Agents, a tool that enables AI agents like ChatGPT or Claude to execute trading operations and manage cryptocurrency payments autonomously. Using the x402 protocol, these agents can pay for digital services without human intervention, a step toward an economy where AI agents act as new financial players.
Visa and OpenAI announce a collaboration to manage transactions initiated by artificial intelligence. While promising advanced security, questions arise about consumer trust and control in an increasingly autonomous ecosystem.
A multidisciplinary team has developed Latent Context Language Models (LCLM), a new architecture that compresses input context up to 16x without significant loss of accuracy. This solves the computational bottleneck of context windows in LLMs, especially in autonomous agents.
The automation of AI research is advancing rapidly. An analysis by Import AI suggests we may be close to a tipping point where machines improve their own algorithms. This raises fundamental questions about control, safety, and the future of work.
AI is transforming modern battlefields: autonomous drones, satellites with onboard processing, and software agents that write military code. This article analyzes the latest advances, their strategic implications, and the risks of an arms race without clear rules.
Google has announced two major innovations at its I/O 2026 conference: Gemini Omni, a multimodal model that integrates text, image, audio, and video in real time, and Gemini 3.5 Flash, an optimized version for speed and low cost. Both products redefine the AI landscape, directly competing with OpenAI and consolidating Google's leadership in the sector.
NVIDIA has unveiled a set of new physical AI skills at CVPR that enable researchers and developers to automate complex workflows for autonomous vehicles, robotics, and computer vision. These skills, integrated with the Cosmos 3 model and NVIDIA libraries, promise to accelerate research by bridging the gap between AI models and complete workflows.
Google has filed a lawsuit against a Chinese cybercrime network that used its own artificial intelligence Gemini to create phishing websites and send millions of fraudulent text messages. The operation, dubbed 'Outsider Enterprise,' offered a phishing-as-a-service (PhaaS) via Telegram, allowing third parties to launch scams without technical knowledge.
Microsoft has introduced ASSERT, an open-source framework that transforms written requirements into automated tests for AI agents. With only 1% of organizations evaluating agents before production, this tool aims to standardize behavioral validation in enterprise environments.
The European Union has ordered Meta to allow WhatsApp to interoperate with third-party artificial intelligence chatbots, such as ChatGPT. The measure, based on the Digital Markets Act (DMA), aims to foster competition, but Meta calls it 'regulatory overreach' that would benefit big tech companies like OpenAI.
Prompt injection is the number one threat for LLM-based applications. A team from Berkeley has developed StruQ and SecAlign, two fine-tuning-based defenses that separate instructions from data and optimize preferences, reducing optimized attacks to less than 15% success.