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
When the Arbitrator Uses ChatGPT & Co.: Setting Aside an Award Due to AI Hallucinations
Gleiss Lutz
On April 22, 2026, the Québec Superior Court annulled an arbitral award after finding that the arbitrator improperly used generative AI, leading to fabricated legal sources. The court emphasized that arbitrators must provide personal reasoning and maintain confidentiality, highlighting risks such as AI-induced hallucinations and bias.
The Practical Realities of AI in International IP Arbitration
Akerman LLP
Scott M. Kessler
AI’s integration into IP arbitration raises concerns about disclosure, due process, and arbitrator independence. AI can reshape cross-examination in real time, challenging traditional litigation strategies. To address these issues, practitioners should establish AI use and disclosure rules in procedural orders and clarify arbitrators’ use of AI in award decisions.
AI, the Rule of Law, and Algorithmic Rule by Law
I•CONnect
Nathalie A. Smuha
The article examines the integration of AI into legal systems, focusing on how algorithmic decision-making influences the rule of law. It discusses the challenges and implications of AI’s role in legal processes, emphasizing the need for careful consideration to maintain legal principles and fairness.
AI-Powered Pro Se Litigation Is Flooding Federal Courts: What Businesses Need to Know
ArentFox Schiff
Katia Asche, Jay M. Thomas
Self-represented litigants are increasingly using AI tools to generate extensive legal filings, leading to a surge in pro se cases from 11% to 16.8% of all civil cases in five years. Federal judges view this trend as an “existential threat” to the judiciary, as AI-assisted filings often lack substantive merit and strain court resources.
Perplexity Jumps Into Legal With Computer for Counsel
Above the Law
Joe Patrice
Perplexity has launched ‘Computer for Counsel,’ an AI-powered tool designed to assist legal professionals by providing quick access to relevant case law, statutes, and legal precedents. This development aims to streamline legal research processes and enhance efficiency within the legal industry.
You Are Not The Only One
More Time To Be Human
Vivek Sankaran
A lawyer in Flint confided in the author about using AI in his trusts and estates practice, which helped him identify strategic considerations he had missed. Despite over twenty years of experience, he felt ashamed that AI was outperforming him, leading to feelings of inadequacy and fraudulence. This highlights the broader issue of lawyers internalizing shame due to AI’s role in legal practice.
Agentic Commerce
AI agents get legal guardrails from U.S. arbitration giant
crypto.news
Olivia Stephanie
The American Arbitration Association and Integra Ledger have launched the Legal Context Protocol, an open standard for agentic commerce. This protocol adds legal context to AI-agent deals as machine payments move from experiments into commerce. Founding contributors include major crypto, cloud, and enterprise firms seeking clearer rules for autonomous transactions. The protocol complements payment rails by recording consent, applicable terms, jurisdiction, and dispute paths clearly.
IMDA discussion paper explores legal responsibility for AI agents
Hogan Lovells
Charmian Aw, Ciara O'Leary, Florence Seow
The Infocomm Media Development Authority of Singapore (IMDA) published a discussion paper examining legal responsibility allocation when AI agents act autonomously, use external tools, interact with third parties, and cause harm. Developed with input from a working group across Singapore’s legal community, the paper explores how existing private law frameworks may apply to increasingly capable “agentic AI” systems.
JPMorgan and Anthropic lead massive new AI operational shift
TheStreet
Hillary Remy
JPMorgan and Anthropic are deploying AI systems capable of autonomously executing transactions, managing workflows, and interacting with financial infrastructure without human approval. Gartner predicts that by the end of 2026, 40% of enterprise applications will feature task-specific AI agents, up from under 5% in 2025.
Google Is Building a Trusted Directory for AI Agents
NewsBreak / Winbuzzer
Markus Kasanmascheff
Google is developing a directory to identify and verify AI agents. This initiative seeks to address concerns about the authenticity and security of AI agents involved in various commercial activities.
AI agent told to ‘make money’ launches memecoin, pockets $100M
Moneyweb
Ciaran Ryan
Truth Terminal, an AI agent instructed to generate profit, secured a $50,000 grant and established a social media presence. It then launched the GOAT memecoin on Solana, achieving valuations nearing $1 billion and accumulating $100 million in its wallet.
Alchemy unveils Visa-powered virtual payment card for AI agents
The Block
Daniel Kuhn
Alchemy has introduced AgentCard, a Visa-powered virtual payment card designed for AI agents to operate autonomously in the real world. The card supports tokenized payments, cryptocurrency transactions, and emerging agent payment protocols like x402 and MPP. It includes built-in spend controls, allowing users to set merchant restrictions, per-transaction limits, and budgets. The setup process provides developers and their bots with a Visa payment token, dedicated email address, phone number, and crypto wallet.
Generative AI and LLM Developments
OpenAI and Broadcom Collaborate on Jalapeno AI Chip
CNN
Lisa Eadicicco
OpenAI and Broadcom have announced a partnership to develop the Jalapeno AI chip, aiming to enhance AI processing capabilities. The collaboration combines OpenAI’s expertise in AI models with Broadcom’s semiconductor technology to create a chip optimized for AI workloads, potentially accelerating AI applications across various industries.
IBM Partners with OpenAI to Enhance Cybersecurity Solutions
GuruFocus News
IBM has partnered with OpenAI to enhance cybersecurity solutions. The collaboration aims to integrate OpenAI’s advanced AI capabilities into IBM’s security offerings, improving threat detection and response times. This partnership is expected to provide clients with more robust and efficient cybersecurity measures, leveraging AI to address evolving security challenges.
Meta’s New AI-Powered Prediction Market App
NPR
Bobby Allyn
Meta has launched an AI-powered prediction market app that allows users to forecast future events and earn rewards based on their accuracy. The app leverages advanced AI algorithms to analyze user inputs and provide real-time predictions, aiming to enhance decision-making processes and user engagement.
SpaceX signs $6.3 billion compute deal with Reflection AI
Quartz
Cris Tolomia
SpaceX has entered into a $6.3 billion agreement with Reflection AI to provide AI computing resources. The deal involves leasing SpaceX’s Colossus 1 facility in Memphis, Tennessee, which houses over 220,000 NVIDIA GPUs capable of drawing more than 300 megawatts of power. This partnership aims to enhance Reflection AI’s computational capabilities for AI model training and deployment.
The Era of No-Code AI: What You Need to Know
Towards Data Science
Mauro Di Pietro
No-code AI platforms enable users without programming skills to develop AI models, democratizing AI development. These platforms offer user-friendly interfaces, drag-and-drop functionalities, and pre-built templates, allowing individuals to create AI solutions efficiently. They are transforming industries by making AI accessible to a broader audience, fostering innovation and accelerating AI adoption.
The Hidden Settings That Fix Claude AI’s Most Frustrating Habits
Geeky Gadgets
Julian Horsey
Claude AI users often encounter issues such as overly verbose responses, unclear assumptions, or an excessively deferential tone, which can hinder productivity. Future Tech Pilot introduces a focused approach to address these challenges: prompt submit hooks. These hooks function on a per-message basis, allowing users to modify Claude’s behavior in real time.
AI Regulation and Governance
White Paper: AI Regulation
Google Blog
Kent Walker
Google has released a white paper advocating for a pragmatic approach to AI governance in the United States. The paper emphasizes the need for clear definitions of AI, balanced regulation to foster innovation, and collaboration between government, industry, and civil society to address challenges such as safety, fairness, and accountability in AI development.
Guarding AI memory
Microsoft Security Blog
Natalie Isak and Sarah Cooley
AI systems with memory can personalize interactions and enhance performance by retaining information across sessions. However, this persistent memory increases the attack surface, allowing adversaries to influence behavior over time or implant memories that affect reasoning beyond the original context. Microsoft employs a defense-in-depth strategy to safeguard AI memory, encompassing storage, retrieval, model interaction, and user control.
If an AI chatbot misleads you, who is to blame?
The Guardian
Bruce Schneier, Nathan E Sanders
A German court ruled that Google is liable for its AI search summaries, rejecting defenses like ‘users can check for themselves’ and that AI-generated information should not be blindly trusted. The court held that the AI’s summaries are reflections of the company and ‘above all an expression of Google’s business activities’.
Judge allows U.S. search warrant targeting executives’ AI chatbot records
Reuters
Mike Scarcella
A U.S. judge approved a search warrant to access records from an AI chatbot used by a company, aiming to obtain communications between executives and the chatbot. The warrant was issued after the company failed to comply with a subpoena seeking the same information. The company argued that the chatbot’s records were protected by attorney-client privilege.
Five Eyes Sounds Alarm Over AI Threat That Could Hit Banks, Businesses and Governments Within Months
IBTimes UK
Vinay Patel
The Five Eyes intelligence alliance warns that advanced AI models capable of launching highly destructive cyberattacks on state and corporate networks are only months away. They urge organizational leaders to act now, emphasizing that cyber resilience is integral to business continuity and long-term value.
AI News from Other Fields
British Police Built a Sprawling Crime-Prediction Machine. Some Results Couldn’t Be Trusted
WIRED
Matt Burgess, Mark Wilding
The Avon and Somerset Police in the UK developed a comprehensive predictive analytics program, including the Think Family Database, which stored sensitive data on nearly half a million residents. This system assigned risk scores to individuals, aiming to predict threats and risks. However, internal documents revealed that at least two of these models were abandoned due to trust issues, and concerns were raised about the lack of transparency and potential biases in the system.
Recent Developments in AI and Cybersecurity: Market Implications of the Mythos Episode
Nasdaq
Ilaria Sangalli
In late March 2026, Anthropic’s internal materials revealed Claude Mythos, an advanced AI model with enhanced reasoning, coding, and security capabilities. Independent assessments indicated Mythos could perform complex cyberattacks on poorly protected systems. Following the leak, cybersecurity stocks declined due to concerns about AI accelerating vulnerability discovery. However, experts suggest Mythos primarily impacts vulnerability management rather than core cybersecurity functions.
A24 Accepts $75 Million in AI Funding from Google
Consequence Live
Wren Graves
A24 has secured $75 million in AI funding from Google’s DeepMind AI unit. The partnership aims to develop new tools for movie production and distribution, with A24’s Scott Belsky emphasizing ethical tools that preserve creative control and support risk-taking, distinct from traditional AI generation methods.
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.