Machine Learning

OpenAI’s GPT-5.2 Release Redefines Professional AI Capabilities OpenAI has released GPT-5.2, its latest and most advanced AI model, specifically engineered for professional knowledge work and enhancing economic value. The new model demonstrates significant improvements in tasks like creating spreadsheets, building presentations, and writing code. GPT-5.2 also showcases enhanced capabilities in image perception, understanding long contexts, and managing complex, multi-step projects. According to OpenAI, the model sets a new standard across numerous benchmarks, outperforming industry professionals in specified knowledge work tasks across 44 different occupations. This release is a strategic response to recent advancements from competitors like Google’s Gemini and Anthropic’s Claude, intensifying the AI landscape competition. The model is being rolled out in tiers, starting with paid plans in ChatGPT, and is available to all developers via the API, marking a significant step in OpenAI’s strategy to accelerate AI adoption in creative, analytical, and operational roles. ...

OpenAI Warns Future AI Models Pose ‘High’ Cybersecurity Risks OpenAI has issued a stark warning that its upcoming frontier AI models could present ‘High’ level cybersecurity risks. The company stated that these advanced models might be capable of automating the discovery and exploitation of significant cyber vulnerabilities, including the development of functional zero-day exploits. To mitigate these potential threats, OpenAI is implementing a ‘defense-in-depth’ strategy. This includes training its models to refuse harmful requests while remaining useful for defensive and educational purposes. The company is also establishing a Frontier Risk Council, an advisory board of cybersecurity experts, to help manage these advanced AI risks. Furthermore, OpenAI is developing a trusted access program to provide enhanced AI capabilities to qualified cybersecurity professionals for defensive operations. ...

Okibo’s EG7+ Robot: AI-Powered Drywall and Painting Automation Arrives in the US Okibo has officially launched its EG7+ robot in the United States, introducing an autonomous, AI-guided system for commercial painting and drywall finishing. This battery-powered robot is engineered for high-wall projects, boasting a reach of up to 24 feet for sanding, painting, and achieving a level 4 finish. A standout feature is its patented AI-driven 3D scanning and real-time modeling algorithm, which enables navigation without external markers or BIM tools. The EG7+ can cover an average of 1,000 square feet per hour, significantly boosting productivity and reducing physical strain on construction workers. The robot made its debut at the Federal and AEC Innovation Day hosted by Hensel Phelps, where it garnered positive feedback from industry leaders. ...

New AI System Enhances At-Home Balance Training for Physical Therapy Researchers at the University of Michigan have developed a new machine learning model to support home-based balance training for physical therapy patients. The system utilizes data from four wearable sensors to predict how a physical therapist would assess a patient’s performance on balance exercises. This AI-driven feedback can help patients make faster progress and maintain their abilities after their formal therapy sessions have concluded. The model was developed by training convolutional neural networks with sensor data and evaluations from physical therapists, achieving nearly 90% accuracy in predicting patient balance ratings. The technology aims to make balance training more accessible, particularly for those in rural areas, and to assist physical therapists in making healthcare decisions. ...

EU Investigates Meta’s WhatsApp AI for Antitrust Violations The European Commission has launched an antitrust investigation into Meta, focusing on concerns that its policies for WhatsApp could unfairly promote its own AI assistant. The probe follows complaints from smaller companies alleging that upcoming changes to WhatsApp’s access rules might hinder competing AI chatbots from reaching users. EU officials are considering interim measures to address these concerns while the investigation proceeds. This action is separate from, but complementary to, existing regulations under the Digital Markets Act. Italy’s antitrust authority is also conducting a parallel investigation. ...

MIT Unveils Dynamic Reasoning Method to Boost LLM Efficiency Researchers at MIT have engineered a more intelligent method for large language models (LLMs) to allocate computational resources during reasoning, significantly increasing their efficiency. This new technique enables LLMs to dynamically adjust the amount of computation used based on the complexity of a given question. This approach marks a departure from common methods that assign a fixed computational budget to every problem, which often wastes resources on simple queries and fails to solve more complex ones. By enhancing the reliability and efficiency of LLMs for complex reasoning, this development could lower the energy consumption of generative AI systems and enable their use in more critical, time-sensitive applications. The research is being presented at the Conference on Neural Information Processing Systems. ...

China’s AgiBot A2 Humanoid Robot Completes Record-Breaking 106-Kilometer Walk Setting a new world record for humanoid robotics, the AgiBot A2 has successfully walked 106 kilometers (66 miles) from Suzhou to Shanghai. Developed by Suzhou-based Zhiyuan Robotics, the robot completed the three-day journey as a test of its endurance and engineering. The AgiBot A2 showcased its advanced capabilities by navigating real-world environments and autonomously swapping its own batteries during the trek. While the company has not specified the frequency of battery swaps or the degree of human oversight, it confirmed the robot navigated using its onboard sensors and algorithms. This long-distance achievement marks a significant milestone in humanoid technology and environmental adaptability. ...

MIT Unveils BoltzGen: A Generative AI for Designing Therapeutic Molecules In a major leap for drug discovery, scientists at MIT have introduced BoltzGen, a generative AI model capable of creating novel protein binders from scratch. This innovative tool is designed to address diseases that have long been considered difficult or impossible to treat, marking a significant shift from predictive biology to generative engineering in medicine. How BoltzGen Unifies Design and Prediction BoltzGen stands out by unifying protein design and structure prediction into a single, powerful process. The model incorporates specific constraints to ensure that the molecules it generates are not only novel but also functional and stable, adhering to the fundamental laws of physics and chemistry. Unlike previous models that were often limited to specific protein types, BoltzGen is engineered as a more general-purpose tool, capable of creating a much wider range of therapeutic molecules. ...

Anthropic Unveils Claude Opus 4.5, Setting New AI Benchmarks Anthropic has launched Claude Opus 4.5, its most advanced and powerful AI model to date. Now accessible via the Anthropic API and major cloud platforms, Opus 4.5 reportedly establishes a new industry standard for complex tasks, including coding, AI agent development, and multi-step computer operations. The model also demonstrates significantly improved performance in everyday business functions like in-depth research and processing documents such as spreadsheets and presentations. Anthropic is positioning this release for enterprise and engineering teams managing sophisticated projects. Claude Opus 4.5 is available on AWS Bedrock, Google Cloud’s Vertex AI, and Microsoft’s Foundry. ...

Leaked Trump Administration Draft Order Aims to Block State AI Laws A leaked draft executive order from a potential Trump administration reveals a plan to preempt states from enacting their own artificial intelligence laws. The proposed order suggests using the withholding of federal funds, including from the Broadband Equity, Access, and Deployment (BEAD) program, as leverage against states that implement “onerous” AI regulations. The document also outlines the creation of a Department of Justice task force to legally challenge state-level AI legislation. This move is widely seen as an attempt to establish a single federal standard for AI, a position President Trump has previously supported on social media. The proposal has drawn sharp criticism from civil society groups, who argue it represents federal overreach and undermines state-level consumer protection efforts. ...