OpenAI has officially launched GPT-5, marking a significant advancement beyond GPT-4o. This new model focuses on complex reasoning and real-world interaction, offering enhanced capabilities for multi-step problems, a larger context window, and fewer factual errors. It also features advanced native multimodal capabilities, seamlessly integrating text, video, and audio. Meanwhile, cryptocurrency exchange Bybit introduced AI Trading Skills, allowing users to execute trades and manage portfolios using natural language commands through AI assistants. This feature provides full market access via 253 API endpoints and includes robust security measures like testnet practice.
On the ethical front, Grammarly removed its 'Expert Review' AI feature after facing a multimillion-dollar lawsuit. The feature controversially mimicked the writing styles of prominent figures like Stephen King without consent for commercial gain. In a move to secure AI intellectual property, Corvex launched Secure Model Weights, which uses hardware-based Trusted Execution Environments and NVIDIA's Confidential Computing to protect AI model weights on third-party infrastructure. This ensures model weights are cryptographically isolated and only decrypted within the GPU. Furthermore, top AI researchers are making significant career moves, with Mustafa Suleyman, co-founder of DeepMind and Inflection AI, now leading Microsoft's AI division.
Beyond commercial applications, an NIH-funded AI tool now predicts intimate partner violence risk in patients with 88% accuracy, aiming for earlier interventions. Rensselaer Polytechnic Institute's Lally AI Academy is also fostering future talent, offering a 30-day program for students to build AI products, starting April 1, 2026. The Church of Jesus Christ of Latter-day Saints is actively engaging in discussions about AI ethics and regulation, with Elder Gerrit Gong emphasizing safety and proper oversight. Global law firm Akin advises companies on navigating the complex transactional and regulatory landscape of AI, including generative AI and autonomous systems. This period also sees prominent researchers, including Yann LeCun from Meta and Ilya Sutskever from OpenAI, launching new ventures, signaling a dynamic shift in the AI research ecosystem.
Key Takeaways
- OpenAI launched GPT-5 with enhanced reasoning, a larger context window, reduced hallucinations, and native multimodal capabilities.
- Bybit introduced AI Trading Skills, enabling crypto trading with natural language commands via 253 API endpoints and strong security.
- Grammarly removed its 'Expert Review' AI feature due to a multimillion-dollar lawsuit over unauthorized imitation of writers' styles.
- Corvex launched Secure Model Weights, using NVIDIA's Confidential Computing and TEEs to protect AI model weights on third-party infrastructure.
- An NIH-funded AI tool predicts intimate partner violence risk in patients with 88% accuracy using routine medical data.
- Rensselaer Polytechnic Institute's Lally AI Academy offers a 30-day program for students to build AI products, with the first sprint beginning April 1, 2026.
- Prominent AI researchers, including those from Meta and OpenAI, are leaving major tech companies to launch new venture-backed startups.
- Mustafa Suleyman, co-founder of DeepMind and Inflection AI, now leads Microsoft's AI division.
- The Church of Jesus Christ of Latter-day Saints is actively discussing AI ethics and regulation, emphasizing safety and proper oversight.
- Akin, a global law firm, provides legal services for AI companies on transactional, regulatory, and policy matters across multiple regions.
Bybit enables crypto trading with natural language AI
Bybit has launched AI Trading Skills, allowing users to trade cryptocurrencies using natural language commands. This feature uses AI assistants to interpret user instructions and execute trades, manage portfolios, and access market data. With 253 API endpoints, it offers full market access through six modules, covering everything from live prices to advanced features like margin lending. Bybit emphasizes security with safeguards and testnet practice for users, aiming to simplify complex trading operations.
Bybit launches AI Skills for easy crypto trading
Cryptocurrency exchange Bybit has released AI Trading Skills, enabling users to trade and manage portfolios using simple, natural language commands. This feature requires no setup and works with major AI platforms, offering full market access through 253 API endpoints. It includes modules for market intelligence, spot and derivatives trading, and account management. Bybit has implemented strong security measures, including testnet practice and explicit confirmation for live trades, to ensure user safety.
Bybit's AI Skills simplify crypto trading with natural language
Bybit has introduced AI Skills, a new feature that allows users to trade cryptocurrencies using natural language commands with AI assistants. This tool requires no installation and is compatible with various AI platforms, offering access to 253 API endpoints for comprehensive market interaction. The system includes six modules for market intelligence, trading, and account management. Bybit has prioritized security with features like testnet trading and explicit confirmation for all live transactions.
Bybit AI Skills offer zero setup crypto trading
Bybit has launched AI Skills, a feature that allows users to trade crypto using natural language commands through AI assistants. This tool requires no setup and works with major AI platforms, providing access to 253 API endpoints for full market interaction. It covers market intelligence, spot and derivatives trading, and account management. Bybit has implemented robust security measures, including testnet practice and explicit confirmation for live trades, to ensure user safety and control.
Grammarly retracts AI feature after backlash and lawsuit
Grammarly has removed its controversial 'Expert Review' AI feature following widespread criticism and a multimillion-dollar lawsuit. The feature used generative AI to mimic the styles of prominent writers and academics, using their names for commercial gain without consent. Investigative journalist Julia Angwin is leading the class-action lawsuit, arguing this practice is unlawful. Grammarly's parent company, Superhuman, apologized and stated the feature had minimal usage, but will defend against the claims.
Grammarly removes AI writer imitation feature after backlash
Grammarly has disabled its 'Expert Review' AI feature, which imitated the styles of famous writers and academics. The company faces a multimillion-dollar lawsuit for using individuals' identities without permission for commercial purposes. The feature, which provided feedback supposedly from figures like Stephen King and Neil deGrasse Tyson, has drawn criticism from those included. Grammarly's parent company, Superhuman, apologized and is redesigning the feature, while also stating they will defend against the lawsuit.
OpenAI's GPT-5 promises advanced reasoning and multimodal capabilities
OpenAI has officially launched GPT-5, presenting it as a significant leap beyond previous models like GPT-4o. This new AI is designed for complex reasoning and real-world interaction, aiming to be a reliable partner for problem-solving. Key improvements include drastically enhanced reasoning for multi-step problems, a much larger context window for processing extensive data, and a notable reduction in factual errors or 'hallucinations'. GPT-5 also boasts advanced native multimodal capabilities, allowing it to understand and integrate text, video, audio, and other data formats seamlessly.
AI tool predicts intimate partner violence risk in patients
Researchers funded by the NIH have developed an AI tool to help clinicians predict which patients are at risk of intimate partner violence (IPV). This decision support system uses machine learning trained on routine medical data to identify potential risks years before patients might seek help. The AI can detect IPV risk with 88% accuracy, outperforming current screening methods. This tool aims to facilitate timely interventions and improve public health outcomes for millions affected by IPV.
RPI launches AI Academy for students to build products
Rensselaer Polytechnic Institute's Lally School of Management has launched the Lally AI Academy, a 30-day program for students to design, build, and launch AI-powered products. Open to all students regardless of coding background, the academy provides AI tools, cloud infrastructure, and mentorship. Teams receive resources to develop ideas into functional products, gaining practical skills in project management, teamwork, and ethical AI use. The first sprint begins April 1, 2026.
Corvex secures AI model weights with hardware protection
Corvex has launched Secure Model Weights, a new solution that protects valuable AI model weights on third-party infrastructure. Using hardware-based Trusted Execution Environments (TEEs) and NVIDIA's Confidential Computing, it ensures model weights remain cryptographically isolated and are only decrypted within the GPU. This prevents even the infrastructure provider from accessing sensitive intellectual property. The system uses remote attestation and post-quantum key exchange for robust, silicon-level security.
Utah's LDS Church navigates AI ethics and regulation
The Church of Jesus Christ of Latter-day Saints is actively engaging in discussions about the ethical development and regulation of artificial intelligence. Elder Gerrit Gong has spoken about AI's limitations compared to divinity, emphasizing safety and proper regulation. Utah is becoming a hub for AI discussions, with leaders seeking a balance between faith and modernity. The state has faced federal opposition, particularly from the Trump administration, regarding its efforts to regulate AI and big tech companies.
Top AI researchers launch new startups
Prominent AI researchers are leaving major tech companies to found their own startups, backed by significant venture capital. These include pioneers like Yann LeCun, formerly of Meta, and Ilya Sutskever, a co-founder of OpenAI. Fei-Fei Li, known as the 'godmother of AI,' and Mira Murati, former interim CEO of OpenAI, are also launching new ventures. Mustafa Suleyman, co-founder of DeepMind and Inflection AI, has moved to lead Microsoft's AI division. These moves signal a new wave of innovation driven by experienced AI leaders.
AI war games explored in new documentary 'Intelligence Rising'
The documentary 'Intelligence Rising' explores the future of artificial intelligence through a unique war game scenario. Featuring AI leader Marc Warner and other global experts like General Patrick Sanders and philosopher Yuval Noah Harari, the film examines how AI could reshape global power. Directed by Alina Andreicheva, the movie premiered at CPH:DOX 2026. It delves into the creators' concerns about the technology they helped develop and discusses AI's learning process by comparing it to human development.
Akin Law firm advises on AI technologies and regulations
Akin, a global law firm, provides comprehensive legal services for companies and investors in the artificial intelligence sector. Their team advises on a wide range of AI technologies, including generative AI, autonomous systems, and AI hardware. They help clients navigate complex transactional, regulatory, and policy matters across the US, Asia, Europe, and the Middle East. Akin's lawyers possess deep expertise in technology, government, and AI governance, assisting clients with export controls, compliance, and emerging legal challenges.
Sources
- Bybit Introduces AI Trading Skills for Natural-Language Crypto Trades
- Bybit Announces Launche Of AI Skills: Powering AI Agents for Crypto Trading With Zero Setup
- Bybit Launches AI Skills: Powering AI Agents for Crypto Trading With Zero Setup, 253 API Endpoints and Growing
- Bybit Launches AI Skills: Powering AI Agents for Crypto Trading With Zero Setup, 253 API Endpoints and Growing
- Vindicated At Last In My Years-Long Loathing Of Grammarly
- Grammarly removes AI Expert Review feature mimicking writers after backlash
- GPT-5: Tutto ciò che sappiamo sul prossimo modello di OpenAI (2026)
- Researchers develop AI tool to predict patients at risk of intimate partner violence
- RPI's Lally School of Management Launches AI Academy for Student Teams to Build and Launch Products in 30 Days
- Corvex Launches Secure Model Weights, Delivering Hardware-Enforced Protection for AI Inference on Third-Party Infrastructure
- How The Church of Jesus Christ of Latter-day Saints Found Itself in the Battle Over Big Tech
- These AI Researchers Quit To Launch Their Own Startups
- AI War Games in 'Intelligence Rising' Doc, Premiering at CPH:DOX 2026
- Artificial Intelligence
Comments
Please log in to post a comment.