Irvin DeAndrei (Dee) Drummond

291 pages into the "One Big Beautiful Bill," Section 43201(c) creates a federal moratorium preventing states from enforcing any AI regulation - including existing laws on #deepfakes, nonconsensual imagery, and consumer protection. 40 out of 50 state Attorneys General and 140+ organizations opposed this provision. It affects hundreds of existing state laws, not to mention opening the floodgates for major market players to exploit the data they already have. Every day, we are seeing federal policy systematically dismantle the regulatory frameworks that smart, thoughtful people have spent years building. States like California, Colorado, and others have pioneered ethical AI oversight. This bill would eliminate that progress and replace it with nothing. The broader pattern is concerning: from gutting diplomatic expertise at State/USAID to imposing impossible work requirements on food assistance recipients, we're witnessing the deliberate dismantling of institutional knowledge and social infrastructure. For those of us in tech policy, our professional expertise increasingly feels irrelevant when evidence-based governance is abandoned. The burnout is real when you see carefully constructed regulatory frameworks and support systems—pesky things like human rights, civil rights, environmental regulations protecting consumers, healthcare and research funding, I can go on—destroyed overnight. This isn't "political:" it's about whether expertise, institutional knowledge, and evidence-based policy still matter in America. Do you think they do, or are we all just waiting for the crash and hoping to rebuild after? #AIGovernance #DataPrivacy #TechPolicy #RegulatoryPolicy #nokings

6

Claude is my operating system. It's a CLM, Word plugin, assistant and agent all rolled into one. Right now, while I'm answering legal questions, Claude is running in the background restructuring my folder system, adding context where it needs to be, organizing everything under a single file that tells it how to work with me and how to work with legal at Worksome. The CLAUDE.md file. Think of it as the instruction manual for Worksome's legal department and me. Except I didn't write it for humans. The structure is three layers: 1️⃣ The CLAUDE.md file It includes who I am, how legal works, what I care about, how I like information presented, how I want AI to think and handle queries, which folder to go to for which task. 2️⃣ A context document written for AI in each folder Plus the actual knowledge. Ya know - the files and Word docs lawyers love. 3️⃣ Claude has navigation instructions In the CLAUDE.md file and context files within each folder. Baby experiments first. Low stakes. When those work, and only then do I apply it to the things that matter. It's a little scary. Mostly exciting. And I genuinely have no idea what the end state looks like yet. I'm documenting all of this in AI Field Notes from a GC. Next issue is entirely about my CLAUDE.md file and Cowork. If you're building something similar and rethinking where context lives, how your systems are structured, what it means to actually run legal on AI, I'd love to compare notes. (See, I told you I didn't need a CLM or a Word plugin. 😉)

83

26 Comments

AI is useful tool; this post and underlying article underscore the need for validating its resulting output. Other recommendations are honing query skills, as well as tailoring AI generated material to your particular issue. Common sensical, yes, yet time pressures contribute to “as is” submissions. (Thanks, Rouz Tabaddor!)

9

📢 Global Privacy & AI Laws Evolving Fast! Here's a quick rundown of key recent developments you need to know: 👇 California's CIPA & Chat Communications: The Ninth Circuit Court of Appeals has clarified that California's Invasion of Privacy Act (CIPA) Section 631(a) only applies to telephone communications, not internet-based chats. This is a significant ruling for businesses using website chat functionalities, affirming that services like Salesforce operating such chats don't violate this specific section of CIPA. 📞💬 EU Copyright Law & Generative AI: A new study from the European Parliament Research Services (EPRS) sheds light on a critical issue: only works with identifiable human input can gain copyright protection in the EU, leaving AI-generated content in a legal grey area. The study also highlights challenges with AI training on copyrighted data and unclear opt-out mechanisms. Lawmakers are urged to clarify rules to keep pace with generative AI. 🤖✍️🇪🇺 California's New AI Employment Rules: Big news for HR and tech! California is set to enforce new rules prohibiting the use of AI tools in employment decisions if they lead to discrimination based on protected traits like age, gender, or disability. This applies to everything from resume screening to job ads, aiming to ensure fair AI usage and uphold anti-discrimination laws. 💼⚖️🚀 Spanish DPA Fines ASNEF for GDPR Breach: The Spanish data protection authority (AEPD) has fined ASNEF €200,000 for unlawful processing of personal data and failing to respond to an erasure request under GDPR. This underscores the importance of proper data notification, accurate record-keeping, and fully honoring individuals' data rights. A stark reminder for all organizations! 🇪🇸🔒💰 #DataPrivacy #AILaw #GDPR #California #EUCopyright #EmploymentLaw #TechRegulation #LegalUpdate #StayInformed #PrivacyCompliance #DPO #LegalHead

6

On June 25, Connecticut Governor Lamont signed Senator James Maroney’s SB 1295 into law. Connecticut has - again - reset the standard. Here's what you need to know: 1️⃣ The bill makes many significant changes to Connecticut’s existing consumer data privacy law, including modifying its applicability standard, exemptions (including the GLBA entity-level exemption), definitions, consumer rights, privacy notice provisions, data minimization standard, and children’s privacy sections. 2️⃣ The bill significantly modifies the law’s approach to profiling that will impact the use of artificial intelligence in some contexts. This includes requiring impact assessments and providing a new Minnesota-inspired (but different) right to question the result of profiling. We talk about these changes extensively in the below article, including how the changes are impacted/limited by the law's exemptions and how small wording changes have a significant impact on things like the law's applicability to pipeline decisions. 3️⃣ This is an article unlike any we've written before. For each of the changes, we analyze not only what changed but what that change means and how it fits into the larger puzzle with state privacy laws in California, Montana, Colorado, Oregon, Minnesota, and Maryland. The goal of the article is to contextualize the amendments and provide nuance and background so that the reader can understand what's really happening and why. 4️⃣ Finally, perhaps the most significant development is not contained in the bill at all. As part of the budgeting process, Senator Maroney secured an increase in funding for the Attorney General’s office to hire more enforcers. Article co-authored by Shelby Dolen and Marlaina Pinto

83

4 Comments

CPPA building bridges with industry. “We’re going to have a completely open door approach, and philosophy,” said Kemp, who started the role last month. “All voices are going to be heard, and we’re going to try to strike the right balance.” - California Privacy Protection Agency executive director Tom Kemp tells Bloomberg Government that the CPPA will work proactively to improve its relationship with the business community that has repeatedly criticized its rulemaking for overreach. https://lnkd.in/eejDV9Vx

18

Yesterday, the U.S. Supreme Court chose to not to hear the appeal of Thaler v. Perlmutter from the D.C. Court of Appeals. This was a case concerning the copyrightability of art generated by AI. Dr. Thaler (a computer scientist) attempted to register a piece of art that was "created autonomously" by a machine that he named the "Creativity Machine." Both the U.S. Copyright Office and the Trump Administration urged the Supreme Court to not take up Thaler's appeal. The D.C. Court of Appeals decision was based on the fact that the machine was not a human and, therefore, could not possess rights in the work of art. The Court held that human authorship still remains a "bedrock requirement of copyright" ownership. Importantly, the D.C. Court of Appeals, in denying the Creativity Machine ownership in the work of art, did note that "[t]he Copyright Office, in fact, has allowed the registration of works made by human authors who use artificial intelligence." There is a pressing need for refined copyright laws to address ownership issues related to AI-generated art, code, and other work products. However, I agree that the Thaler case was not the appropriate platform for this decision. We need Congress to act first.

16

3 Comments