🪴Thanks for reading AI Governance, Ethics & Leadership! If you enjoyed the post, don’t be shy about it! Like it, add a comment, or a restack!! ⬆️ Use my Linkedin Summary Optimizer Free 🌿 Estimate Your AI Environmental Impact
Last year, I predicted that employees would begin pushing back on these layoffs - which seem unfounded in many instances. My prediction was early - but it’s materializing now.
In one of the first high-profile ageism lawsuits to emerge from its 2025 layoffs, former Senior Director of Monetization Analytics Nicolas Franchet (who spent 13 years at Meta rising through the ranks) was terminated at age 54. Despite a strong performance history, he was suddenly rated a “lowest performer” and cut during a round framed as targeting the bottom 5%. As a senior leader, Franchet was privy to critical internal discussions on promotions, raises, and talent strategy. Because of his seniority, he also actively participated in AI-driven layoff planning. His suit alleges that workers over 40 were 1.5× more likely to be laid off than those under 40, while those over 50 were 2.5× more likely, raising serious questions about how AI-augmented performance systems and restructuring tools may embed or amplify age bias.
This suit comes at a time where the job market is challenging, inflation is up and regulations on AI don’t yet protect employees. Nick’s experience isn’t a one off and Meta isn’t the only company employing these tactics. This suit is likely the first of many.
This week’s Top 5 includes:
Meta Age Discrimination Lawsuit (Nicolas Franchet)
OpenAI Drops Sora & Altman’s Shifting Responsibilities
Big Tech Compensation Shrinks (Salesforce + Meta)
NYC Health + Hospitals Drops Palantir Contract
Claude’s 5 Outages And Service Tiers March 2026
This week’s developments score a 51 on the EvA index.
The EvA index ranks weekly developments on a single 0–100 scale where 0 = maximally exploitative and 100 = maximally accountable, so readers can instantly see whether leaders acted to protect people, rights, and oversight or to prioritize speed, control, or extraction.
1. Meta Age Discrimination Lawsuit (Nicolas Franchet)
Exploitation vs Accountability Index: 30 (Exploitation)
Former Meta Senior Director Nicolas Franchet (13‑year tenure, strong performance history) filed a lawsuit in March 2026 alleging age discrimination in the 2025 layoffs. According to the complaint, workers over 40 were 1.5× more likely to be cut — and those over 50 were 2.5× more likely — after performance ratings were allegedly manipulated to classify older employees as “lowest performers.” Source: Gizmodo.
Impact to AI Governance (AD’s Take)
Shame on Meta for adjusting their performance processes to one less rigorous and one that targeted older workers. What stood out to me here is how quickly subjective or manipulated performance decisions became a vehicle for age discrimination - first enacted by humans, then likely reinforced as a pattern inside AI systems. Both the EU AI Act and NIST AI RMF explicitly flag age‑related bias as a governance risk, but in the U.S., federal protections apply regardless of whether discrimination comes from a manager or an algorithm. That includes the Age Discrimination in Employment Act (ADEA), EEOC enforcement, and relevant provisions of the Civil Rights Act and ADA.
👉 This suit all but guarantees increased scrutiny on how organizations evaluate, rank, and offboard workers especially for anyone over 40.
2. OpenAI Drops Sora & Altman’s Shifting Responsibilities
Exploitation vs Accountability Index: 50 (Ethical Gray Area)
OpenAI shut down its standalone Sora video app/API (and Disney partnership) due to high compute costs and low usage. CEO Sam Altman is reallocating focus to capital raising, data-center builds, and enterprise/AGI deployment, stepping back from direct safety/security oversight. Source: TechCrunch
What AI Governance Insiders Are Saying
What’s interesting is the conversation in governance circles. Here’s a quote from Violeta that captures the sentiment: Every video Sora generated carried a moderation burden that scaled faster than the product itself. Deepfakes of real people within hours. Copyrighted characters within days. The cost of governing that output was growing faster than the revenue it could ever produce. OpenAI didn’t just prioritize enterprise. They exited a product where the governance cost was the business model’s ceiling. - Violeta Klein, CISSP, CEFA
What No One Is Talking About ( AD’s Take)
Not enough people are talking about the significance of the executive shift that occurred as a result of dropping Sora. Sam Altman is specifically stepping back from direct oversight of safety and security to focus on OpenAI’s profitability and compute. That change can be interpreted as strategic but it’s also a deprioritization of existing safety standards - essentially safety teams will report into a head of AGI.
👉 The deprecation of Sora won’t translate to less deepfake videos being created. Consumers will still have to be more vigilant than ever.
3. Big Tech Compensation Shrinks (Salesforce + Meta)
Exploitation vs Accountability Index: 50 (Ethical Gray Area)
Salesforce froze merit raises for director-level and above in 2026, redirecting funds to larger stock/bonus pools for top performers. Meta revived performance-vested stock options for key execs (first since 2012) with aggressive long-term market-cap hurdles.(Source: Business Insider
Implications for Workers & AI Governance (AD’s Take)
AI capital expenditure is now so large, hundreds of billions across the industry, that it’s beginning to crowd out human compensation. This is what I call Phase 2. Phase 1 was AI‑driven layoffs. Phase 2 is compensation compression: raises slow, salary bands tighten, and equity is increasingly reserved for a shrinking tier of “AI‑critical” talent.
As AI systems take on more of the work that once justified high salaries, tech leaders are becoming more aggressive about reallocating compensation toward key AI roles …while squeezing the broader workforce and governance layers.
👉This shift continues to expose the gap in worker protections as AI capabilities grow. U.S. federal labor law was not designed for a world where capital expenditure on compute replaces labor expenditure on people, and where productivity gains accrue to shareholders and AI‑specialized roles rather than the general workforce.
4. NYC Health + Hospitals Drops Palantir Contract
Exploitation vs Accountability Index: 70 (Accountable)
NYC Health + Hospitals (largest U.S. municipal healthcare system) will not renew its ~$4M Palantir contract (expiring October 2026) after public backlash over use of patient health notes for revenue-cycle optimization and broader concerns about data handling by the controversial vendor. (Read More)
Impact to AI Governance & Patients (AD’s Take)
This case exposes a deeper governance failure: patients had no meaningful way to consent to or opt out of having their clinical notes processed by a third‑party analytics platform. Sensitive PHI was already flowing into Palantir’s system before ethical review or public scrutiny caught up. In a public healthcare system where patients cannot choose their provider, the absence of consent removes the only real control individuals have over how their data is used.
👉I call this as a classic “too‑late” governance problem (that could have been avoided): once the data is in motion, governance becomes reactive rather than preventative.
Note: Although the EU AI Act classifies healthcare‑operations AI as high‑risk and the NIST AI RMF emphasizes data governance, privacy, and third‑party risk, the U.S. still lacks enforceable federal AI‑specific regulation for healthcare.
5. Claude’s Outages And Service Tiers March 2026
Exploitation vs Accountability Index: 55 (Ethical Gray Area)
Anthropic’s Claude platform experienced at least FIVE notable outages in March 2026 (including a nearly five‑hour disruption around March 27) driven by demand surges and networking issues. Enterprise surfaces (claude.ai, Claude Code, Opus 4.6) were hit hardest. Government‑tier customers were often less affected, reflecting different reliability guarantees and infrastructure isolation. Source: The New Stack
What AI Governance Insiders Revealed (AD’s Take)
Most people don’t connect an AI outage to AI governance — but they should. Under the EU AI Act and NIST AI RMF, AI service providers are expected to deliver reliability, robustness, and operational transparency. One outage is unusual. Five outages in a single month signals deeper structural fragility.
So I asked the obvious question:
👉 How did government systems stay online while enterprise and consumer tiers collapsed?
The answer is governance‑coded into infrastructure:
Government workloads run on isolated, higher‑redundancy compute
Public‑sector SLAs are stricter, enforceable, and financially punitive
Vendors prioritize uptime where contractual penalties are highest
Enterprise and consumer tiers subsidize reliability for government clients
This isn’t malicious — it’s economic gravity.
But it exposes a truth the industry rarely says out loud:
AI reliability is stratified. The more leverage you have, the more uptime you get.
Startups, SMBs, and everyday consumers absorb the operational risk. Government clients (the ones with the most bargaining power) get the most stable service.
🏁 The Close Out
These five developments paint a coherent but messy picture: the AI industry is deep in a sloppy growth crisis. Explosive demand and massive capital investment are colliding with the realities of infrastructure limits, human incentives, and institutional lag. None of this is random noise. These are the predictable consequences of treating frontier AI as pure infrastructure while governance, reliability, and fairness remain afterthoughts.
Claude’s outages, Meta’s age‑bias lawsuit, NYC Health’s retreat from Palantir, and OpenAI’s safety‑leadership shift all point to the same underlying tension: scaling velocity is outrunning institutional maturity. The systems are getting bigger faster than the guardrails can be built.
And this is where coordinated governance responses matter. If NYC Health drops Palantir, the question isn’t whether one contract ends — it’s whether other NYC agencies are reviewing their own Palantir dependencies. One department’s reversal doesn’t move the needle unless it triggers a broader institutional audit. The same applies across the industry: isolated fixes won’t stabilize a system whose risks are systemic.
The through‑line is simple:
AI is scaling like infrastructure, but we’re governing it like a game of wack-a-mole.
Until we have coordinated responses, these crises won’t be exceptions, they’ll be the standard.
About
AD is a Stanford Certified Engineering Leader & MBA turned AI Governance Leader supporting small and midsize executive teams.
👋 Follow AI Governance, Ethics & Leadership on LinkedIn and X fka Twitter
⬆️ Use my Linkedin Summary Optimizer Free
🌿 Estimate Your AI Environmental Impact
💵 Check Your AI CEO Values (You’ll be surprised)
