🐣 First edition

Shadow AI | Litigation funders back in business | Fake reviews crackdown

Good morning! Welcome to Profiles in Legal - the weekly newsletter helping in-house counsel stay ahead of the curve.

We design each edition to give you the best ROI on your time: legal and regulatory updates that help you spot headwinds early, ask smarter questions, and steer your business more confidently through change.

This week, we’re starting with a quiet revolution already underway inside most companies: AI is making calls on hiring, firing, and promotions, often with zero legal oversight. Our lead story breaks down what’s happening behind the scenes, and what Legal should do next.

BRIEFING ROOM

Do you know where your AI is tonight?

While legal teams are busy drafting AI policies, debating procurement language and sweating over foundation model transparency, AI has quietly walked in through the side door, taken a swivel chair and started making decisions.

🪑 Middle managers are disappearing, says payroll platform Gusto. As businesses flatten, each manager now oversees more direct reports, meaning less capacity for difficult conversations and performance management nuance. Enter: more AI. Less oversight. Resume Builder reports that 1 in 5 managers now use AI to make final calls on promotions, raises, and even firings. Two-thirds of them have had no formal training. So yes, ChatGPT might be your new HR business partner. (Hope it’s read your Code of Conduct.) 

Meanwhile, companies are encouraging “AI fluency” while trimming management layers. The “human in the loop” is increasingly an overworked team lead trying to hit OKRs and keep Slack under control.

🍟 Even approved tools can cause headline-grabbing issues. This week, McDonald’s hit pause on its AI hiring chatbot, used to screen applicants, after 64 million applicant records were exposed - thanks to the password “12345” (saltier than when its drive-thru bot ordered 260 McNuggets and ice cream topped with bacon).

👩‍💼Unapproved tools - aka “shadow AI” are a growing headache. These aren’t just rogue hacks from resourceful interns. Often, it’s senior leaders adopting AI tools at pace, worried about falling behind in the AI arms race. The risks multiply as tools become more agentic (making decisions, not just suggestions) (AI Magazine).

What Legal Can Do This Week

(Besides panic.)

🕵️‍♀️ Ask around: Where is AI already being used day-to-day? Not just what’s on the roadmap - what’s already in motion?

🧭 Recalibrate your radar: Don’t just look at procurement. Look at process and at high risk areas (HR, client proposals). Where are decisions being delegated - and to whom (or what)?

🛠️ Plan for reality, not theory: Governance frameworks are good. But guardrails, training, and escalation pathways for AI-in-use? Even better. Aim for one secure, approved AI tool, blocking the rest, rather than none, which just incentivises shadow workarounds.

RISK RADAR

🇪🇺 The European Commission published its final draft General-Purpose AI Code of Practice, to help providers of general-purpose AI models (think OpenAI, Google, Meta) comply with the EU AI Act, especially safeguards for copyright-protected work, transparency and systemic risks. It doesn’t apply to regular companies using AI for business.  Reuters.

🇬🇧 The UK government has proposed voiding NDAs that prevent staff from speaking out about sexual misconduct, harassment, or discrimination - prompted by the Can’t Buy My Silence campaign (led by Zelda Perkins, ex-Harvey Weinstein aide). If passed, this could limit the use of confidentiality clauses in settlement agreements, making workplace disputes less appealing to resolve quietly and increasing litigation risk. BBC.

⭐️ Stars, stats and spotlight  The UK’s CMA has started enforcing new rules on fake and misleading reviews under the Digital Markets, Competition and Consumers Act 2024. B2C businesses pulling reviews from third-party platforms like Trustpilot or Google are most at risk - often a blind spot for Legal. To avoid risk of fines of up to 10% of annual turnover, start checking for red flags like recycling old reviews for new campaigns; and undisclosed incentivised reviews.

EXPLAINER OF THE WEEK

BACKED TO THE FUTURE

Court of Appeal clears litigation funding workaround post-PACCAR If you breathed a sigh of relief when the Merricks v Mastercard class action fizzled out, you weren’t alone. Momentum for UK competition claims dipped. And after the PACCAR ruling nuked traditional litigation funding agreements, it looked like funders might be out of the game too. Not anymore. Last week, the Court of Appeal ruled that newly structured litigation funding agreements (LFAs) are enforceable - even with caps linked to damages. That’s a green light for funders, and it revives the UK's collective actions pipeline.

🧱 Quick Recap: What Was PACCAR Again? In PACCAR v CAT (2023), the Supreme Court held that LFAs paying funders a cut of damages were really damages-based agreements (DBAs). That made them unenforceable in most collective actions. The workaround? Funders rewrote LFAs to pay themselves a multiple of the funds invested—an input-based model. But potential defendants (Visa, Mastercard, Sony, Apple…) challenged those too.

  What the Court Just Decided

  • Input-based LFAs with caps are fine. A return structured as a multiple of investment plus a cap linked to damages doesn’t convert the LFA into a DBA.

  • Fallback clauses are allowed. If the law changes later and the agreement reverts to a % of damages, that doesn’t invalidate the current structure.

  • Severability is unresolved. The court sidestepped the question of whether unenforceable clauses in older LFAs can simply be cut out.

💡 Our Take: Why This Matters to In-house Counsel

Funding is back and here to stay. Revised LFAs are now court-approved. Funders have legal certainty again and are already regrouping.

Expect a class action rebound. Claims paused post-PACCAR, especially in the Competition Appeal Tribunal, are likely to restart. New ones will follow.

Litigation funding is now part of risk strategy. Whether you’re defending or advising, scrutinising the funding structure should be part of your early due diligence.

HIRING BOARD

💰 Revolut is looking for a Legal Counsel (Regulatory), 4 PQE, remote from various European locations.

IN THE CALENDAR

🇳🇱 Crafty Locals Amsterdam Friday 8 August 2025, 18:00, Amsterdam - informal in-house networking. Sign up.

FROM THE SIDEBAR

👨🏼‍⚖️ “No one should ever know what I’m thinking” - a must-watch interview with the Quinn Emanuel partner who counts Elon Musk, Jay-Z and Mr Beast as clients

🧭 We’d like to be your compass in acing your job, but in case you need a real one

❤️ Who knew there were songs about lawyers in love

Enjoying Profiles in Legal?

Our readers are curious, commercially sharp and allergic to legalese. If that’s you - welcome.

💬 Forward to a fellow innovator in Legal

Too much legal content is dull, jargon-filled or humblebrag. We’re changing that - and building a sharper legal community in the process.

🔗 Subscribe

🪃 Reply to this email with what you think we should cover

📣 Request to advertise with us

© 2025 Profiles in Legal