This human-curated, AI-generated newsletter from the AAA-ICDR Institute and AAAiLab keeps you up on AI news over the past week that is relevant to alternative dispute resolution (ADR).
AI in ADR and Legal Services
Scaling justice: How AI and ADR are reshaping legal access
Thomson Reuters Institute
Maya Markovich
AI-driven platforms are transforming dispute resolution by automating and streamlining processes that have traditionally been slow, costly, and inaccessible for many. By integrating AI with alternative dispute resolution methods, these systems can handle high volumes of cases efficiently, reduce barriers to justice, and offer tailored solutions. The challenge ahead is ensuring these technologies are designed for fairness, inclusivity, and transparency, enabling broader and more equitable access to legal remedies.
Artificial Intelligence and the Modern General Counsel
FTI Consulting
General counsels are increasingly expected to guide their organizations through the adoption of generative AI, balancing innovation with legal and ethical oversight. Despite widespread interest, most legal teams feel unprepared for the risks involved. Success hinges on strategic leadership, cross-functional collaboration, adaptive governance, and a focus on purposeful, risk-aware AI integration—positioning the GC as both a driver of value and a steward of accountability in the AI era.
Is AI Changing the Outside Counsel Relationship for Better or Worse? A New Report Weighs In
Corporate Counsel / Definely
Kathryn Lye
As artificial intelligence gains traction in legal services, many in-house legal teams remain uncertain about their outside firms' use of AI and demand clear, practical benefits rather than flashy tools. Concerns about data security, trust, and usability persist, leading to a cautious approach. Ultimately, law firms that can transparently demonstrate how AI improves collaboration and outcomes will strengthen client relationships, while others risk losing business to in-house teams.
Imagine reviewing six million documents by hand. For help, these lawyers turned to AI
blog.dropbox.com
Jesse Will
Faced with millions of historical property records that were only available as unsearchable images, Santa Clara County collaborated with Stanford’s RegLab to develop a tailored AI system. By fine-tuning an open-source language model and building advanced text-cleaning tools, the team enabled rapid, accurate identification of racial covenants in difficult-to-read documents, dramatically reducing review time and cost while ensuring high accuracy for required legal oversight.
Generative AI and LLM Developments
Surging investment in artificial intelligence is fueling unprecedented growth among startups, with this year’s leading disruptors reaching record-breaking valuations and aggressively expanding through acquisitions. The landscape is rapidly diversifying, as innovative AI-driven business models emerge across sectors like defense, agriculture, transportation, and healthcare. The influx of new companies on the Disruptor 50 list highlights how embracing generative AI is now essential for staying competitive and relevant in the tech industry.
Meta's $15 Billion Scale Deal Could Leave Gig Workers Behind
TIME
Billy Perrigo
Meta plans a $15 billion investment to acquire nearly half of Scale AI, positioning Scale’s CEO to lead Meta’s new AI initiative aimed at advanced intelligence. While this move could boost Meta’s competitiveness and benefit Scale’s leadership and investors, the company’s global gig workers—crucial to AI training—are expected to see no improvement in low pay or working conditions. The deal may also restrict Scale’s services to Meta’s rivals, intensifying industry competition.
Apple may be the only tech company getting AI right, actually
CNN
Allison Morrow
Apple’s latest developer conference underwhelmed those hoping for major AI breakthroughs, as anticipated upgrades to Siri and other AI features were notably absent. The company is prioritizing reliability over rushing out generative AI tools, reflecting broader industry skepticism about current AI capabilities. Investors are pressuring Apple to accelerate its AI strategy, potentially through acquisitions, as competitors advance and expectations for transformative AI continue to rise.
The Illusion of Thinking: Understanding the Strengths and Limitations of Reasoning Models via the Lens of Problem Complexity
Apple Machine Learning Research
Parshin Shojaee, et al.
Researchers examined how advanced language models that generate detailed reasoning steps actually perform when faced with increasingly complex problems. By testing these models in controlled puzzle environments, they found that both Large Reasoning Models (LRMs) and standard models struggle and eventually fail as complexity rises. Surprisingly, LRMs do not always outperform standard models, and their reasoning often becomes inconsistent, revealing fundamental limits in current AI reasoning abilities.
What "Working" Means in the Era of AI Apps
Andreessen Horowitz
Olivia Moore, Marc Andrusko
Generative AI startups are achieving revenue milestones and fundraising rounds at unprecedented speeds, with both enterprise and consumer companies surpassing traditional growth benchmarks. The performance gap between average and top startups is widening, especially as consumer-focused AI businesses now generate significant revenue quickly, often tied to new model launches. Success increasingly depends on rapid product development and strong user retention, making speed and engagement critical competitive advantages.
AMD Sees AI Inference Growing 80% Annually
Yahoo Finance
Moz Farooque
AMD forecasts that AI inferencing—using AI models to generate results—will soon surpass training as the main force behind data center growth, driven by a surge in specialized models across industries. The company unveiled new GPUs promising significant performance and cost advantages over competitors, with major tech firms already planning upgrades. AMD’s expanding market share and strong customer endorsements signal robust prospects as demand for efficient AI hardware intensifies.
AI Regulation and Policymaking
What If China Wins the AI Race?
Foreign Affairs
Adam Segal, Sebastian Elbaum
U.S. policymakers have pursued export controls and light regulation to sustain America’s AI lead. However, breakthroughs by Chinese firms such as DeepSeek show the gap narrowing. China is advancing AI integration in manufacturing and skirting export restrictions, so dominance is not assured. To prepare for a future where U.S. models finish second, Washington should develop richer evaluation metrics. It should also standardize APIs to reduce lock-in, build adjudication layers, and craft nuanced data-sharing guidelines to ensure continued AI benefits.
Hogan Lovells UK AI Summit 2025: Investment, innovation, and the future of AI regulation
JD Supra / Hogan Lovells
Liam Pape
Industry leaders at the UK AI Summit emphasized the nation's strong position in global AI, highlighting both opportunities and challenges in investment, regulation, and responsible innovation. Key themes included the need for agile, globally interoperable legal frameworks, evolving intellectual property and privacy concerns, and the critical role of boards and investors in AI governance. The event underscored a shift toward practical strategies for scaling and safely deploying AI across sectors.
All civil servants in England and Wales to get AI training
The Guardian
Rowena Mason
Starting this autumn, over 400,000 civil servants in England and Wales will receive hands-on AI training aimed at boosting productivity and streamlining tasks, as the government simultaneously reduces workforce size. AI tools are already being piloted to analyze consultations and correspondence, saving significant time and costs. While officials emphasize modernization, concerns remain about potential errors and bias in automated decision-making without sufficient oversight.
Microsoft Plans to Rank AI Models by Safety
PYMNTS.com
Microsoft will begin rating AI models by safety, adding this metric to its Azure Foundry platform alongside quality, cost, and speed. This move aims to help customers make more informed decisions amid concerns over AI risks, particularly as autonomous systems become more prevalent. The change reflects a broader industry shift toward transparency and trust, especially in regulated sectors like banking, where understanding and managing AI-driven risks is increasingly critical.
South Korea's Evolving AI Regulations
Stimson Center
Seungmin (Helen) Lee
South Korea enacted the world’s first national AI Basic Act, aiming to boost AI innovation while addressing societal risks like deepfakes and unethical AI use. The law establishes new oversight bodies and mandates transparency but is criticized for vague definitions, limited enforcement, and not directly tackling key issues such as political or sexual deepfakes. Its effectiveness will depend on forthcoming regulations and whether enforcement mechanisms are ready by its 2026 start date.
AI News from Other Fields
I Want AI to Plagiarize My Work (opinion)
Inside Higher Ed
Kevin Frazier
The author argues that academic concerns over plagiarism and attribution are inhibiting the transformative potential of generative AI in knowledge creation and dissemination. Rather than prioritizing personal recognition, scholars should embrace AI’s ability to synthesize, remix, and amplify ideas, enabling broader accessibility and impact. The focus should shift from individual credit to maximizing the utility and reach of knowledge, accepting that AI-driven “plagiarism” can ultimately serve the greater good.
Google's AI search features are killing traffic to publishers
TechCrunch
Rebecca Bellan
The rise of AI-powered search tools like Google’s AI Overviews and chatbots is sharply reducing web traffic to news publishers by providing direct answers, often sourced from their content without credit or clicks. This threatens the sustainability of quality journalism, prompting publishers to seek new revenue models, such as licensing deals with tech firms or partnerships to share ad revenue when AI systems use their material.
Amazon is about to be flooded with AI-generated video ads
The Verge
Jess Weatherbed
Amazon has launched an upgraded AI-powered Video Generator tool for US sellers, enabling quick creation of realistic, customizable video ads with features like motion, multiple scenes, and music. Sellers can generate ads in minutes, even from existing footage or images, making professional-quality advertising more accessible and affordable. This move is poised to accelerate the adoption of AI-generated ads across Amazon’s marketplace, potentially transforming how products are marketed online.
Inside the Secret Meeting Where Mathematicians Struggled to Outsmart AI
Scientific American
Lyndie Chiou
A new generation of AI reasoning models, like OpenAI’s o4-mini, has demonstrated remarkable skill in solving advanced mathematical problems, even rivaling expert mathematicians and outperforming previous AI systems. At a recent gathering, top mathematicians struggled to create questions the AI couldn’t solve, highlighting both the potential and the unsettling speed of progress. This shift suggests mathematicians may soon collaborate with, rather than compete against, AI in mathematical discovery.
Lockheed Martin launches 'AI Fight Club' to test algorithms for warfare
SpaceNews
Sandra Erwin
Lockheed Martin is launching "AI Fight Club," a virtual platform where companies—large and small—can pit their AI algorithms against simulated military scenarios to prove their effectiveness for defense use. By opening access beyond traditional defense contractors, Lockheed aims to surface innovative AI solutions, vet real capabilities, and connect top-performing teams with Pentagon opportunities, while ensuring participants' intellectual property is protected and results meet Department of Defense standards.
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.