Lead story
N.D. Cal. Magistrate Rejects AI-Specific Changes to Model Protective Order in Litton v. Roblox
In Litton v. Roblox Corp., No. 25-cv-03088 (N.D. Cal. May 27, 2026), Magistrate Judge Peter H. Kang rejected Defendants' proposed changes to Section 7.1 of the Model Protective Order that would have addressed the use of Confidential/Highly Confidential Information with AI tools. Judge Kang ruled that his existing Standing Order for Civil Cases already contains multiple provisions governing AI tool use by lawyers, experts, and parties, and Defendants failed to show sufficient cause to modify the Model Protective Order.
- What happened
- In Litton v. Roblox Corp., No. 25-cv-03088 (N.D. Cal. May 27, 2026), Magistrate Judge Peter H. Kang rejected Defendants' proposed changes to Section 7.1 of the Model Protective Order that would have addressed the use of Confidential/Highly Confidential Information with AI tools. Judge Kang ruled that his existing Standing Order for Civil Cases already contains multiple provisions governing AI tool use by lawyers, experts, and parties, and Defendants failed to show sufficient cause to modify the Model Protective Order.
- Why it matters
- This ruling signals that courts with existing AI standing orders may reject party attempts to craft bespoke AI provisions in protective orders, preferring to apply uniform judicial AI rules instead. Litigation teams should review local standing orders for AI provisions before proposing protective-order modifications, and should document AI prompt usage to satisfy judicial record-keeping requirements.
- Watch
- This is a magistrate judge order in one district; other courts may rule differently. The case is ongoing. Published June 15, 2026.
SourceseDiscovery Todayapp.minerva26.comedrm.net
Briefing item
Anthropic Suspends Claude Fable 5 and Mythos 5 Following US Government Export-Control Directive
Anthropic suspended access to Claude Fable 5 and Claude Mythos 5 on June 12, 2026, following a US government export-control directive citing national security concerns. The directive required Anthropic to suspend all access to these models by foreign nationals, inside or outside the United States, including foreign-national Anthropic employees. Anthropic stated it must 'abruptly disable Fable 5 and Mythos 5 for all our customers to ensure compliance,' while access to all other Anthropic models remains unaffected.
- What happened
- Anthropic suspended access to Claude Fable 5 and Claude Mythos 5 on June 12, 2026, following a US government export-control directive citing national security concerns. The directive required Anthropic to suspend all access to these models by foreign nationals, inside or outside the United States, including foreign-national Anthropic employees. Anthropic stated it must 'abruptly disable Fable 5 and Mythos 5 for all our customers to ensure compliance,' while access to all other Anthropic models remains unaffected.
- Why it matters
- For legal-tech teams, this is a concrete example of how AI model availability can be disrupted by government action rather than commercial decisions. It strengthens the case for model-agnostic architectures that allow switching between providers, and highlights the importance of supply-chain resilience when building AI-dependent legal workflows.
- Watch
- Anthropic's public statement does not specify the technical basis for the national-security concern, and cybersecurity concerns around Fable 5 were reported separately. Published June 15, 2026.
SourcesLegal IT Insideranthropic.com
Briefing item
HaystackID Launches COMET for Mobile Compliance Oversight and Showcases AI Solutions at LegalTechTalk London
HaystackID announced the launch of COMET (Compliance Oversight for Mobile Electronic Transmissions), a new tool for monitoring and managing electronic communications across mobile and collaboration platforms. The company showcased COMET alongside its enhanced AI solutions at LegalTechTalk London, positioning the product within its broader CoreFlex and Core Intelligence AI platform ecosystem.
- What happened
- HaystackID announced the launch of COMET (Compliance Oversight for Mobile Electronic Transmissions), a new tool for monitoring and managing electronic communications across mobile and collaboration platforms. The company showcased COMET alongside its enhanced AI solutions at LegalTechTalk London, positioning the product within its broader CoreFlex and Core Intelligence AI platform ecosystem.
- Why it matters
- Mobile and ephemeral messaging continues to be one of the hardest data sources for eDiscovery collection and compliance. A dedicated compliance-oversight tool for mobile transmissions addresses a real gap in many organizations' preservation and collection workflows, particularly for regulated industries where chat and messaging data must be captured for litigation holds and regulatory requests.
- Watch
- Product details are based on HaystackID's announcement and event coverage; technical specifications and pricing were not fully disclosed. Published June 15, 2026.
Sourceshaystackid.com
Briefing item
EU AI Act High-Risk Enforcement Deadline Arrives August 2 — eDiscovery Tools Likely in Scope
The EU AI Act's high-risk AI system provisions take full effect on August 2, 2026, across all 27 EU member states. Annex III, Section 8 of the Act covers AI systems used to 'research and interpret facts and the law and to apply the law to a concrete set of facts' — language that likely encompasses AI-powered document review, predictive coding, and ML-driven relevance classification used in eDiscovery platforms. Fines reach up to EUR 35 million or 7% of global annual revenue.
- What happened
- The EU AI Act's high-risk AI system provisions take full effect on August 2, 2026, across all 27 EU member states. Annex III, Section 8 of the Act covers AI systems used to 'research and interpret facts and the law and to apply the law to a concrete set of facts' — language that likely encompasses AI-powered document review, predictive coding, and ML-driven relevance classification used in eDiscovery platforms. Fines reach up to EUR 35 million or 7% of global annual revenue.
- Why it matters
- eDiscovery platforms using ML for document classification — including Relativity's aiR for Review, Everlaw's Deep Dive, and DISCO's Cecilia AI — may fall within the high-risk classification for litigation teams operating in or with EU data. Vendors and legal teams should assess compliance obligations including documentation, transparency, human oversight, and accuracy requirements well before the August 2 deadline.
- Watch
- The high-risk filter mechanism (significant-risk-of-harm test) has not been tested in court for legal-tech tools. Vendor positioning on this question is still evolving. Published April 2, 2026; deadline context remains current.
Sourcesblog.platinumids.com
Briefing item
KLDiscovery's AI Case Explorer Targets Earlier Insight in Early Case Assessment
KLDiscovery published a best-practices guide on using AI to improve early case assessment (ECA), centered on its AI Case Explorer feature within the Nebula platform. The article argues that AI can help legal teams identify the people, events, evidence, and evolving facts that shape case strategy earlier in the matter lifecycle, giving teams more time to refine their approach as new data emerges.
- What happened
- KLDiscovery published a best-practices guide on using AI to improve early case assessment (ECA), centered on its AI Case Explorer feature within the Nebula platform. The article argues that AI can help legal teams identify the people, events, evidence, and evolving facts that shape case strategy earlier in the matter lifecycle, giving teams more time to refine their approach as new data emerges.
- Why it matters
- Early case assessment remains one of the most time-sensitive and under-automated phases of the eDiscovery workflow. Tools that surface key signals before full-scale processing and review can materially reduce the cost and duration of the discovery lifecycle. However, teams should evaluate whether AI-driven ECA tools integrate with their existing review platforms and whether the analysis is auditable for later stages.
- Watch
- This is vendor-sponsored content published on eDiscovery Today; KLDiscovery is an Educational Partner and sponsor. The article describes product capabilities based on vendor claims. Published June 15, 2026.
SourceseDiscovery Todaykldiscovery.com
Educational only. Not legal advice. Review sources and matter-specific obligations before relying on any workflow decision.