ZwillGen

The European Commission has released its first substantive guidance on classifying AI systems as high-risk under the EU AI Act. The guidance, published to the Commission's website, was released in three parts covering general principles, Annex I use cases, and Annex III use cases. Read together, the guidance goes a long way in clarifying the Act's application to high risk systems and will aid those preparing for compliance. A few notable points: • For general-purpose AI (GPAI) systems, contractual clauses that exclude high-risk uses are not enough on their own to avoid high-risk classification. If a provider's marketing or product positioning still promotes the system as having broad functionality across a generality of contexts, and does not consistently limit its use, the system will be classified as high-risk. • Agentic systems should be assessed as a whole as a single AI system, rather than based on their individual components. Where multiple components or linked actions together serve a high-risk purpose, the combined system should be classified as high-risk, even if individual components would qualify for certain exceptions on their own. • Having a human reviewer review system outputs before they take effect does not change a system's classification. Human oversight is a compliance requirement for high-risk systems, but does not otherwise change a system's intended purpose relative to assessing whether it falls into an Annex III use case. • The exception for systems that perform only narrow procedural or preparatory tasks is to be construed narrowly. The controlling test is whether the system materially influences the outcome of a decision. AI systems that make value judgements, including scoring, ranking, and labeling inputs as more or less useful, do not qualify for the exception. • Under Annex III, the high-risk recruitment use case includes activities that occur before a candidate has submitted a formal application. AI systems used for candidate sourcing, searching candidate databases, and targeted job advertising can all fall within scope. Written by Andrew Eichen & Brenda Leong #EUAIAct #ArtificialIntelligence #AI #EuropeanUnion #EU https://lnkd.in/e4RG9P5b

The ICO’s 2026 guidance makes clear that UK tracking compliance reaches well beyond traditional cookies, with regulators focusing on a broader range of storage and access technologies and holding organizations to stricter standards around consent, transparency, and user choice. Kady Hammer, Zach J., and Melissa Maalouf discuss these developments and what organizations should consider: #DataPrivacy #UK #CookieCompliance #AdTech https://lnkd.in/ecXvEVsC

The FTC announced it has begun enforcing the TAKE IT DOWN Act (TIDA) and launched a new portal for complaints regarding platforms’ compliance with requests to remove nonconsensual intimate imagery. https://lnkd.in/exkQ2QiJ

From Rachel Miller of ZwillGen: "Businesses and investors are increasingly including AI-specific representations and warranties in commercial contracts and agreements, reassessing longstanding data strategies and sharpening their focus on protecting the rights in data that parties provide to one another." 🔗 https://lnkd.in/gzBQaP_k

It was a beautiful Saturday on the National Mall, and across the country as we participated in the Lawyers Have Heart 5K Run/Walk benefiting the American Heart Association. Team ZG showed up in D.C., with remote racers also logging miles in Virginia, Chicago, and Washington State. #LawyersHaveHeart #AmericanHeartAssociation #ZwillGen

AI has made every technology and data deal more complicated and raises new questions around IP ownership, data monetization and business models. I wrote about what lawyers and business teams are working through in this new landscape.

In a recent Cybersecurity Law & Strategy article, Rachel Miller explains how the surge in AI investment is forcing technology and data companies to rethink their commercial contracts. #Contracts #ArtificialIntelligence #AI #LegalNews https://lnkd.in/e_Ji-GR7

Ken Dreifach and other speakers discussed the challenges ad tech and data companies face when creating and offering health-related audiences at their panel earlier today at The NAI Summit. Thank you to his fellow panelists for their perspectives and audience members for attending! Ben Chapman (Swoop), Jessica Hoffman (PulsePoint), & Robert E. (DeepIntent) #HealthPrivacy #HealthAdvertising #AdTech

Shutterstock will pay $35 million to settle FTC allegations that the company charged consumers without informed consent and made subscription cancellations unnecessarily difficult. The proposed order also includes requirements related to disclosures, consent, and cancellation processes. https://lnkd.in/eZqC5gDD