Microsoft, OpenAI, and Anthropic lead the way in AI development and implementation

Several companies are making strides in AI development and implementation. Debevoise & Plimpton is creating an AI model called STAAR, which provides clients with direct access to documents drafted by the firm's lawyers on AI risk. The model is in beta mode and costs $150,000 per year.

Researchers have introduced Enactive AI, a conceptual framework for enterprise and industry reasoning, site-level decision support, and execution feedback. This framework includes four complementary roles and aims to expand the frontier of AI from model capability to system-aware action.

Cloudflare has announced improvements to its AI Search tool, including a new pricing model and the ability to index and search across multiple data sources. The tool provides a single public endpoint for searching and allows users to add custom domains.

The intersection of AI and bookselling is growing, with author Scott Brown sharing his experience selling four books to AI companies. This incident highlights the increasing importance of AI in various industries.

In the realm of research, a systematic literature review of Explainable AI (XAI) in the context of the EU Right to Explanation has been presented. The review highlights gaps in the interdisciplinary synthesis of the current regulatory framework and identifies areas for further research.

Researchers have also proposed EviGraph, an architecture for autonomous research agents that generates hypotheses, executes experiments, and drafts manuscripts. This system aims to address the problem of unsupported claims and inconsistencies in research outputs.

On the security front, Chinese AI models have been found with stolen US large language model watermarks, raising concerns about intellectual property theft. This incident highlights the need for greater security measures to protect AI technologies.

Qualys has announced new capabilities in TotalAI to empower organizations to discover, test, monitor, and govern enterprise AI risk. The tool provides enterprise CISOs with robust AI governance and risk management capabilities.

The Nairobi Securities Exchange plans to launch East Africa's first artificial intelligence-focused exchange-traded fund (ETF) before the end of 2026. The fund will give investors local access to global AI leaders, including Microsoft, OpenAI, and Anthropic.

Key Takeaways

* Debevoise & Plimpton is creating an AI model called STAAR, which costs $150,000 per year and provides clients with direct access to documents drafted by the firm's lawyers on AI risk. * Researchers have introduced Enactive AI, a conceptual framework for enterprise and industry reasoning, site-level decision support, and execution feedback. * Cloudflare has improved its AI Search tool with a new pricing model and the ability to index and search across multiple data sources. * The intersection of AI and bookselling is growing, with author Scott Brown selling four books to AI companies. * A systematic literature review of Explainable AI (XAI) in the context of the EU Right to Explanation has been presented. * Researchers have proposed EviGraph, an architecture for autonomous research agents that generates hypotheses, executes experiments, and drafts manuscripts. * Chinese AI models have been found with stolen US large language model watermarks, raising concerns about intellectual property theft. * Qualys has announced new capabilities in TotalAI to empower organizations to discover, test, monitor, and govern enterprise AI risk. * The Nairobi Securities Exchange plans to launch East Africa's first artificial intelligence-focused exchange-traded fund (ETF) before the end of 2026, giving investors local access to global AI leaders, including Microsoft, OpenAI, and Anthropic.

Debevoise builds AI model with partner loyalty

Debevoise & Plimpton is creating an AI model called STAAR that gives clients direct access to documents drafted by the firm's lawyers on AI risk. The model is in beta mode and costs $150,000 per year. The firm believes its lockstep culture, which encourages collaboration, will help it build AI products. STAAR is built on the Legora legal AI platform and provides clients with a self-help portal to access previous work, including policies, documents, and training videos.

Debevoise builds AI model with partner loyalty

Enactive AI: A Decision-Centric Architecture

Researchers introduce Enactive AI, a conceptual framework for enterprise and industry reasoning, site-level decision support, and execution feedback. The framework includes four complementary roles: Organizational World, Site World, Schema Intelligence, and Enactive Decision Cycle. Enactive AI aims to expand the frontier of AI from model capability to system-aware action.

Cloudflare AI Search improves developer experience

Cloudflare has announced several improvements to its AI Search tool, including a new pricing model and the ability to index and search across multiple data sources. The tool provides a single public endpoint for searching and allows users to add custom domains. Cloudflare also plans to release a plugin for its AI Search tool.

Bookselling in the Age of AI

Author Scott Brown shares his experience selling four books to AI companies, adding $439 to his gross sales. The incident highlights the growing intersection of AI and bookselling.

Explainable AI for the EU Right to Explanation

Researchers present a systematic literature review of Explainable AI (XAI) in the context of the EU Right to Explanation. The review highlights gaps in the interdisciplinary synthesis of the current regulatory framework and identifies areas for further research.

EviGraph: Autonomous Research Agents

Researchers propose EviGraph, an architecture for autonomous research agents that generates hypotheses, executes experiments, and drafts manuscripts. The system aims to address the problem of unsupported claims and inconsistencies in research outputs.

Chinese AI models found with stolen US watermarks

Chinese AI models have been found with stolen US large language model watermarks, raising concerns about intellectual property theft. The incident highlights the need for greater security measures to protect AI technologies.