This human-curated, AI-generated newsletter from the American Arbitration Association’s AAA-ICDR Institute keeps you up on AI news over the past week that is relevant to dispute resolution and to legal services more broadly.
AI in ADR and Legal Services
Beyond Memorials: Will AI Re-engineer Arbitration’s Written Phase?
Osborne Clarke
AI is increasingly integrated into international arbitration, with parties and legal professionals using tools to search, classify, and summarize documents, generate chronologies, and identify data patterns. The technology possesses the potential to fundamentally transform the written phase of arbitration by replacing traditional sequential memorials with shared, LLM-generated syntheses of issues and evidence. This shift aims to improve procedural efficiency while maintaining the integrity of human advocacy.
Artificial Intelligence in mediation: Opportunities and challenges
The New Times (Rwanda)
Janvier Kajyibwami
AI tools support mediators by structuring processes, analyzing information, and managing communication, which reduces administrative burdens in complex disputes. These systems use machine learning to categorize and summarize vast amounts of evidence, allowing for data-driven negotiation strategies. Despite benefits in efficiency and accessibility, AI integration faces risks such as algorithmic errors, transparency issues, and potential socioeconomic inequality regarding access to technology.
AI judges: a possibility?
Law Society of Ireland Gazette
Oisin Suttle
The integration of AI into judicial systems is transitioning from abstract academic theory to a practical field of safety research. While the Irish Judicial Council warns against relying on AI for legal analysis and the EU AI Act classifies such systems as high risk, various international courts are already experimenting with AI tools for case management, judgment drafting, and research support.
The Ethics of the Co-Pilot: An Update from Brazil and Beyond
Kluwer Mediation Blog
Andrea Maia
The Brazilian National Council of Justice approved Resolution 615/2025, which updates the framework for AI in the judiciary by mandating human supervision and prohibiting systems that obstruct human review. International frameworks, such as the IBA Mediation Committee’s 2025 guidelines and the Silicon Valley Arbitration & Mediation Center’s principles, provide standards for AI use in mediation. These developments prioritize human control while private mediation remains an evolving practice area.
AI ethics bill for courts moves ahead in California Assembly committee
Daily Journal
Malcolm Maclachlan
The California Assembly Judiciary Committee unanimously approved SB 574, a bill introducing new safeguards for AI use by legal professionals. The legislation mandates that attorneys verify all case citations to prevent reliance on fabricated AI-generated content and prohibits entering confidential information into public AI systems. The bill also requires judicial officers and arbitrators to disclose reliance on generative AI when drafting legal documents or orders.
UChicago Law School AI strategy bans phones, laptops in class for 1st year students
CBS News
Sara Tenenbaum
The University of Chicago Law School is prohibiting first-year students from using laptops, phones, and tablets in classrooms to curb AI dependence. This policy is part of a broader pedagogical strategy aimed at fostering independent critical thinking. The school will simultaneously teach students the ethical and effective use of AI for future legal practice while implementing AI-resilient assessments and emphasizing essential human legal skills.
Agentic Commerce
All Rise: Internet Court For AI Agents Is In Session
Forbes
Nina Bambysheva
Autonomous AI agents now perform tasks such as shopping, negotiating terms, and executing payments. As these systems scale, they face risks of errors and inter-agent disputes. To address these conflicts, developers are implementing blockchain-based frameworks and AI-driven jury systems. These tools allow agents to establish formal contracts and resolve disagreements through automated evidence evaluation, providing a mechanism for dispute resolution in agentic commerce.
What new rules for agentic AI in payments might look like
Pinsent Masons
David Tilbury
Agentic AI systems autonomously navigate e-commerce to browse, monitor pricing, and complete transactions on behalf of users. Current payment regulations, particularly strong customer authentication requirements, lack provisions for these autonomous interactions. Regulatory focus is shifting toward establishing new consent and authorization frameworks, while clarifying liability allocation between consumers, developers, payment service providers, and merchants to address risks involving errors, fraud, and system hallucinations.
Agentic AI Liability Fuels Issues Reaching Beyond the Law’s Edge
Bloomberg Law
Mark Kelley
Agentic AI systems conduct transactions and manage tasks autonomously, operating with minimal human input. Businesses are building infrastructure to support agents that research, negotiate, and execute purchases. These autonomous processes require established rules for agent interactions, pricing, and authority. Deployers must address potential liability for unauthorized or unintended transactions, necessitating contractual safeguards and robust verification mechanisms to ensure accountability within evolving agentic commerce workflows.
Agentic Commerce: Why AI Agents Need an Open Web
Open Data Science
Agentic commerce involves AI agents that autonomously research, compare, and execute purchases for users. Current infrastructure relies on specific protocols like the Agentic Commerce Protocol or Universal Commerce Protocol to facilitate these transactions. Because merchant fraud detection systems often block automated activity, agents require broader, open standards to successfully navigate and complete checkouts across diverse online retail platforms beyond those currently providing direct integration.
Generative AI and LLM Developments
GPT-5.6 Sol vs Fable 5
Geeky Gadgets
Julian Horsey
GPT-5.6 Sol and Fable 5 offer distinct performance capabilities for varying project requirements. GPT-5.6 Sol emphasizes speed, cost-efficiency, and token efficiency, making it ideal for high-volume, execution-heavy tasks. In contrast, Fable 5 provides superior strategic thinking and creative output, serving as the preferred choice for complex, high-stakes initiatives.
Anthropic releases improved Claude models that can control your computer
SiliconANGLE
Kyt Dotson
Anthropic introduced a public beta of a new computer use capability for the Claude 3.5 Sonnet model on October 22, 2024. This technology allows the AI to perceive and interact with computer interfaces by viewing screenshots, moving a cursor, clicking buttons, and typing text.
State of AI Infrastructure report overview
Google Cloud Blog
Drew Bradstock
Enterprise AI is transitioning from simple conversational models to agentic systems capable of executing complex, autonomous tasks. Research reveals that 83% of organizations require infrastructure upgrades to support production-grade agentic AI. Scaling these workloads presents challenges, including an inference tax and operational complexity. To address these demands, organizations are adopting fluid compute strategies, unified data layers, and hybrid multicloud architectures to improve performance and cost-efficiency.
AI is becoming a bargain hunter’s market, with a few luxury models on top
The Register
Thomas Claburn
AI inference pricing has split into low-cost commodity services and premium frontier models. While costs for standard AI tasks have plummeted—with some estimates showing a drop to $0.40 per million tokens—frontier models remain expensive due to complex agentic workflows, long context requirements, and metered billing. Businesses now face significantly higher operational costs, often shifting toward multi-model strategies to balance performance and expenses.
Open Data for Agents
Hugging Face Blog
NVIDIA and Hugging Face have introduced an open data initiative to support the development and evaluation of AI agents. This effort provides high-quality, openly licensed datasets featuring tool-use traces, multi-turn dialogs, and environment logs to assist developers in building autonomous workflows. These standardized resources facilitate better benchmarking, improve model reproducibility, and help reduce legal friction for researchers and product pilots creating complex AI systems.
AI Regulation and Policymaking
AI 2040: Plan A
AI Futures Project
Daniel Kokotajlo et al.
Plan A proposes a strategic roadmap to delay the development of superintelligent AI until 2040, mitigating existential risks associated with rapid, unchecked AI advancement. This vision includes an international agreement between the U.S. and China to prioritize research transparency, diffuse AI capabilities across global actors, and establish compute limits. These measures create a controlled environment that permits safe, sustainable technological progress while maintaining human control.
Pritzker signs Illinois AI regulation bill
St. Louis Business Journal / Capital News Illinois
Maggie Dougherty
Illinois Governor JB Pritzker signed the Artificial Intelligence Safety Measures Act, which mandates transparency, risk assessment, and third-party audits for large-scale AI developers. The law requires companies to report potential safety incidents within 24 hours and includes whistleblower protections.
Lawmakers probe growing use of Chinese AI models in U.S. companies
CNBC
Kai Nicol-Schwarz
A House Committee investigation is currently examining the risks associated with the increasing adoption of Chinese AI models by U.S. businesses. Companies are turning to these models due to their cost-effectiveness, high levels of customization, and privacy features compared to some domestic alternatives. The inquiry reflects federal concerns regarding the national security implications of integrating foreign-developed software into the American technology infrastructure.
AI accountability agenda: US senator unveils package of bills to curb tech’s harms
The Guardian / MSN
Sanya Mansoor
Senator Ed Markey introduced a legislative package labeled the AI accountability agenda to address risks associated with unregulated AI. The proposal includes federal certification requirements for datacenters through the Federal Communications Commission to mitigate environmental and energy impacts. Additional measures within the broader strategy target workplace surveillance, discriminatory algorithms, and protections for children.
Inside the alternative playbook to AI regulation
Axios / MSN
Maria Curi
AI companies and government entities are increasingly collaborating on rules for cutting-edge AI technology. While this public cooperation suggests success, it obscures a chaotic behind-the-scenes scramble to influence policy. Organizations use a strategy of reverse federalism, lobbying for specific state-level laws to create a more unified regulatory environment across the country.
AI News from Other Fields
RIP, Free Internet
BOYSCLUB
Deana Burke
Cloudflare's new Monetization Gateway lets websites charge AI agents for access, with instant stablecoin micropayments replacing the ad-based model. AI crawlers now make up over half of bot traffic, and agents are unable to view ads, so Cloudflare is repositioning the web's economics.
Patreon Blocks Crawlers From Stealing Creators’ Work for AI Training
404 Media
Samantha Cole
Patreon partnered with Cloudflare to block known AI training crawlers at the network level across all posts. This measure prevents AI companies from scraping creators’ work to train models without permission. While these crawlers are blocked, search and discovery bots remain active to ensure content visibility. The initiative prioritizes creator demands for consent, credit, and compensation regarding how their work is used by AI developers.
NYC Public Schools Puts Technology Through Its Paces With New AI Guidance
Daily News / Chalkbeat
Lizzie Walsh
New York City Public Schools released preliminary guidelines for AI use, establishing a framework to govern technology integration across the nation’s largest school district. The policy uses a traffic-light system to categorize approved versus restricted AI tools.
Reddit is using LLMs to solve a problem LLMs largely created
TechCrunch
Amanda Silberling
Reddit employs large language models to combat a surge in spam and coordinated bot activity, much of which is generated by similar AI technology. By identifying subtle patterns of fake behavior and artificial hype, these tools allow the platform to block 23 million spam views daily. This initiative resulted in a 20% reduction in user exposure to spam between January and March.
Advancing health care AI through ethics, evidence and equity
American Medical Association
The American Medical Association establishes a framework for the responsible use of AI in healthcare, prioritizing ethical principles, rigorous clinical evidence, and health equity. The guidelines emphasize designing AI systems that address meaningful clinical goals, maintain professional medical standards, and ensure accountability. Effective implementation requires transparent communication with patients, robust algorithmic oversight, and active mitigation of biases in training data to prevent healthcare disparities.
The AAA-ICDR and AAAiLab do not endorse, control, or assume responsibility for the accuracy, availability, or content of any summaries, articles, or external sites or links referenced in this newsletter. Neither the AAA-ICDR nor AAAiLab are compensated or otherwise incentivized by any third parties for including content in this newsletter. All copyrights, trademarks, and other content referenced in the articles or article summaries remain the property of their respective owners. If you believe we have linked to material that infringes your copyright rights, please contact us at institute@adr.org, and we will review and address the matter promptly. This newsletter does not constitute legal advice.
The AAA-ICDR respects your privacy. If you no longer wish to receive these emails, you may unsubscribe at any time by clicking the link below.
We also do not share your personal information without your consent. For more information, please refer to the AAA-ICDR’s full privacy policy.