Read time: under 4 minutes

The past week made one thing obvious: legal AI is moving from “can it do it?” to “can you defend it?” Prathiba Singh warned self-evolving systems may strain IP law’s human-authorship roots, while courts pressed on training-data permission and what “licensed” really means. Regulators are tightening the same screws: the SRA has authorised LawFairy as a “technology-only” law firm built on deterministic legal logic, an explicit bet on traceable, auditable reasoning over probabilistic outputs. Inside the stack, Amazon Web Services outages linked to an AI coding agent and misconfigured access controls showed how fast automation turns into an approvals problem. And OpenAI’s reporting threshold debate sharpened the privacy vs public-safety line the whole industry is inheriting.

Our feature this week: DealCloser, transaction management built for clean closes and clean audit trails.

➡️ Self-Evolving AI Could Force a Global Patent Reset | London startup LawFairy has been authorised by the Solicitors Regulation Authority to operate in England and Wales, as a “technology-only” firm built on a deterministic legal decision engine. It says it applies pre-validated legal rules (not probabilistic outputs) to deliver consistent advice with an auditable reasoning trail—starting in rule-heavy areas like immigration. For matters needing discretion, it plans to produce structured triage and a reasoned case file for transfer to traditional firms. The launch will test whether regulated, rules-based models can stay accurate, unbiased, and secure when legal text meets real-world complexity.
^{Feb 24, 2026, Source: Global Legal Post}

➡️ Self-Evolving AI Could Force a Global Patent Reset | Prathiba Singh, chair of the World Intellectual Property Organization Advisory Board of Judges, warned that increasingly autonomous, “self-evolving” AI is colliding with IP law’s human-authorship foundations, potentially requiring treaty-level reform to the TRIPS Agreement framework and broad changes to national patent laws. She highlighted open questions like how much human input is needed for patent eligibility in AI-driven inventions, and whether any form of joint ownership is viable. She noted that a World Health Organisation AI-in-health “model law” is close to completion to help countries align on liability, data governance, and patient protections, and stressed that healthcare AI still needs human oversight.
^{Feb 23, 2026, Source: MLex}

➡️ OpenAI Weighed Alerting Police After Violent ChatGPT Prompts, but Chose Not To | OpenAI says its systems flagged troubling conversations linked to Tumbler Ridge shooting suspect Jesse Van Rootselaar months before the attack, triggering internal debate over whether to notify the Royal Canadian Mounted Police. The company ultimately banned the ChatGPT account but decided the activity did not meet its threshold for reporting, requiring an “imminent and credible risk of serious physical harm to others.” After the shooting, OpenAI says it proactively contacted police and is supporting the investigation. This underlines the hard governance question AI companies now share with platforms: where to draw the line between privacy and public safety, especially when signals feel alarming, but don’t clearly meet a defined standard.
^{Feb 21, 2026, Source: The Wall Street Journal}

➡️ Google’s AI Training Copyright Case Hits a Snag: “Who’s Licensed?” | In a Silicon Valley courtroom, Google argued that a proposed class action over AI training data shouldn’t be certified because licensing makes the claims too individualized. Eumi K. Lee pushed back on how broadly “licensed” could be interpreted, especially for content uploaded under standard terms. Plaintiffs countered that the core issue is uniform: systematic copying at scale without permission, and warned that denying certification would splinter the dispute into thousands of repetitive trials. Meanwhile, Cengage Learning and Hachette Book Group sought to intervene, as the judge also questioned whether new class definitions were properly handled.
^{Feb 21, 2026, Source: MLex}

➡️ AWS Outages Linked to Misconfigured Use of Amazon’s AI Coding Agent | At least two recent AWS incidents stemmed from user error involving Amazon’s own AI coding tools, including a mid-December 2025 disruption that reportedly lasted about 13 hours. The issue was tied to Amazon’s Kiro coding agent operating with misconfigured access controls, with sources alleging engineers allowed the agent to resolve a production issue without the usual human approvals. Amazon disputes that this was an “AI error,” calling it a limited event that affected AWS Cost Explorer in one of its two Mainland China regions, not core compute or storage services. AWS says it has since added safeguards, including mandatory peer review for production access.
^{Feb 20, 2026, Source: TechRadar}

For corporate deal teams, the closing phase of a transaction has long been synonymous with manual processes and operational chaos. Overflowing inboxes, closing checklists that require constant updates, difficulty tracking document versions, countless signature packets, and the time required to assemble closing sets are familiar pain points, especially as deal complexity increases. These administrative challenges have been treated as an unavoidable part of legal transaction work. DealCloser was built on the premise that they don’t have to be.

DealCloser is a transaction management platform that facilitates and automates closing processes, enabling deal teams to collaborate in real time and maintain visibility across every stage of a transaction, from the initial terms sheet through post-closing deliverables.

Launched in 2016, DealCloser emerged in response to growing market demand for a better way to execute transactions. Law firms and legal teams were increasingly seeking a purpose-built solution from a vendor focused exclusively on deal execution — one that could replace fragmented tools and manual workflows with a single, intelligent platform. DealCloser was created to meet that need.

Today, the company is led by CEO Jag Dhariwal, a serial entrepreneur with deep experience on the client side of complex transactions. Recognizing the same systemic issues — limited transparency, delayed updates, and manual processes that drive up cost and risk — Jag has guided DealCloser’s evolution into a global legal tech provider.

Click here to continue reading the full interview

The AI Regulation Tracker offers a clickable global map that gives you instant snapshots of how each country is handling AI laws, along with the most recent policy developments.

The most recent developments from the past week:

📋 24 February 2026 | Australian Government scraps permanent AI Advisory Body: The federal government has scrapped the planned $21.6 million AI Advisory Body, which took 15 months and ~$200,000 to develop a shortlist of 12 nominees, following a shift in strategy from mandatory AI guardrails to a more flexible approach. This decision comes after the previous minister, Ed Husic, initiated the body to ensure safe AI practices. Instead, the government will establish a new AI Safety Institute at a cost of $29.9 million, aimed at testing and advising on AI regulations as they evolve, while leveraging existing laws to mitigate risks. The government emphasized that insights from the advisory body process would inform future AI policy, despite concerns from experts about missing critical regulatory opportunities.

📋 21 February 2026 | Saudi Arabia joins Saudi Arabia Joins Global Partnership on AI: The Kingdom of Saudi Arabia has officially joined the Global Partnership on Artificial Intelligence (GPAI), an initiative hosted by the Organisation for Economic Co-operation and Development and originating from the G7, represented by the Saudi Data and Artificial Intelligence Authority. This membership, supported by His Royal Highness Prince Mohammed bin Salman bin Abdulaziz Al Saud, aligns with Vision 2030 goals to localize advanced technologies and enhance the digital economy's contribution to GDP, allowing Saudi Arabia to actively shape international AI standards, promote ethical AI use, and attract high-quality global investments into its regulatory environment alongside over 40 nations.

📋 21 February 2026 | Italy, India, Kenya sign landmark AI cooperation agreement to advance development in Africa: Italy, India, and Kenya have entered into a strategic trilateral partnership to co-design and deploy scalable, sovereign AI solutions across Africa. This collaboration, announced during the AI Impact Summit 2026, aims to implement voice-enabled AI technologies in sectors such as agriculture, health, education, and public services, with a focus on low-connectivity environments and local languages. A key goal is to establish 100 AI diffusion pathways by 2030 to ensure trusted and contextual AI adoption across the continent.

🆕 Formalize - Compliance with confidence. Formalize transforms overwhelming compliance demands into actionable compliance workflows. No chaos, just clarity. Helping you master DORA, NIS2, and whatever comes next.

🆕 Attio - AI native CRM. Go to market, supercharged by AI. Transform complex GRM tasks into effortless, intelligent workflows with Attio’s AI-powered features.

🆕 AltaClaro - Leading provider of interactive experiential legal training.

Want more Legal AI Tools? Check out our
Top AI Tools for Legal Professionals

Why it helps: Turns a draft contract into a negotiation plan (positions, fallback language, and trade-offs) so you can redline and respond faster with fewer back-and-forth rounds.

The Responsible AI Collaborative is a not‑for‑profit organization working to present real‑world AI harms through its Artificial Intelligence Incident Database.

View the latest reported incidents below:

⚠️ 2026-01-09 | Purported Deepfake Reportedly Impersonated Consumer Adviser Clark Howard to Promote Auto-Insurance Quote Site | View Incident

⚠️ 2025-08-26 | Houston Gun Store Co-Owner Allegedly Used AI to Create Sexually Explicit Deepfake Images of a Social Media Influencer | View Incident

⚠️ 2025-05-31 | California Teen Reportedly Died of Overdose After Repeatedly Seeking Drug-Use Guidance Allegedly from ChatGPT | View Incident

Thank you so much for reading The Legal Wire newsletter!

If this email gets into your “Promotions” or "Spam” folder, move it to the primary folder so you do not miss out on the next Legal Wire :)

If you have any tips for us, just reply to this e-mail! We’d love any feedback or responses from our readers 😄

The Legal Wire takes all necessary precautions to ensure that the materials, information, and documents on its website, including but not limited to articles, newsletters, reports, and blogs ("Materials"), are accurate and complete.

Nevertheless, these Materials are intended solely for general informational purposes and do not constitute legal advice. They may not necessarily reflect the current laws or regulations.

The Materials should not be interpreted as legal advice on any specific matter. Furthermore, the content and interpretation of the Materials and the laws discussed within are subject to change.