LinkedIn's "correction", Vibe Coding risks & The Cognitive Residual

Hello H.A.I.R. Community,

It has been 22 days since my last email. That is terrible, and I apologise.

However, it has been for a good reason. I have been deep in the trenches of algorithmic accountability, specifically regarding the "visibility collapse" many women are experiencing on LinkedIn.

Today’s edition is a catch-up. I want to share the latest on the LinkedIn bias investigation, summarise two critical governance risks I’ve recently uncovered (Vibe Coding and the "Cognitive Residual"), and share where you can hear more from me over the next few weeks.

Let’s get into it.

LinkedIn Responded. They Think They Refuted My Research.

I Believe They Just Confirmed It.

If you have been following my posts, you know we have been investigating whether the LinkedIn algorithm suppresses women’s voices. LinkedIn’s Head of Responsible AI recently published a blog post to "clarify" how the feed works.

I have analysed their response. Far from debunking the issue, it inadvertently confirms the exact mechanism of Proxy Bias I identified in my initial report.

Here is the breakdown of why their explanation fails the fairness test:

• The Denial: They spent most of the post denying they use "gender" as a variable. I agree. I never claimed the code said if gender == female. That would be Direct Discrimination. My argument has always been about Indirect Discrimination via proxies.

• The Admission: They explicitly listed the signals they do optimise for: "Position," "Industry," and "Activity."

• The Problem: Men are historically overrepresented in high-visibility industries (Tech/Finance) and senior roles. By optimising for these signals without a fairness constraint, the system systematically amplifies men. Furthermore, the "Activity" signal favours "agentic" (male-coded) linguistic patterns over "communal" ones.

In the UK, this is the textbook definition of Indirect Discrimination under the Equality Act 2010. In the EU, this is a Systemic Risk under the Digital Services Act (DSA).

The Action: Analysis is important, but action is essential. I am supporting the petition "Calling for Fair Visibility for All on LinkedIn," demanding an independent equity audit.Why Does This Happen? It's Not Just 'Algorithm Aversion'

The "Vibe Coding" Trap: When HR becomes the AI Provider

While the LinkedIn debate rages on, a new trend has emerged in Talent Acquisition: "Vibe Coding."

This is the wave of agentic AI tools that allow non-engineers to "build" apps using natural language prompts. I’ve seen recruiters with zero coding experience build automated offer generators and interview feedback carousels in minutes.

It looks like innovation. It smells like liability.

In my recent detailed newsletter on this topic, I broke down why this is a massive GRC headache:

1. The Red Line: The moment a tool is used for recruitment or worker management, it is classified as High-Risk AI under the EU AI Act.

2. You Are The Provider: If your team "vibe codes" a tool, your organisation is legally the "Provider" (manufacturer). You inherit the conformity assessments, logging, and monitoring duties usually reserved for tech companies.

3. Shadow AI: We are moving from Shadow IT to Shadow AI. These tools often lack authentication or store API keys in plain text, creating GDPR nightmares.

My advice: "Build is liability." Encourage innovation, but wrap it in a governance framework.

The Policy Vacuum: Who Owns Your "Digital Twin"?

I also recently explored a concept I call the "Cognitive Residual."

As we start building AI "digital twins" of expert employees to capture their knowledge and style, we run into a profound legal conflict.

• The Conflict: Companies want to own the "asset" (the model of how an employee thinks). Employees have rights to their "identity" (GDPR and Right of Publicity).

• The Gap: Standard "work-for-hire" contracts govern outputs (reports, code), not cognitive personas.

This is the Asset-Identity Duality, and currently, there is zero policy for it. We need to move from an extractive model to a licensed model.

Listen & Watch: Recent Appearances

If you want to dive deeper into these topics, I have been making the rounds on the airwaves.

🎙️ HR Means Business with Mervyn Dinnen I joined Mervyn to discuss "Navigating the Hype and Reality." We covered the difference between AI capability and competence, and why vendor due diligence is your first line of defence. Listen here

🎥 From AI Hype to Audit-Ready HR In this episode, I shared my "guardrails-first" checklist. We discussed why you should avoid LLMs for CV screening decisions but embrace them for personalisation, and how to speak to your board about AI risk without the jargon. Watch the episode

Upcoming: Where to Find Me

I will be speaking at several key events in December. I’d love to see you there.

📅 9th December: AI in TA - The 2026 Landscape I’m joining a panel with Michael Blakley (Equitas), Khyati Sundaram (Applied), and Jamie Betts (Neurosight) to look at the future of recruitment tech. Save your spot

📅 10th December: Algorithmic Bias Deep Dive Hosted by the EWMD Network. This is a 90-minute deep dive into the LinkedIn algorithm changes and the economic impact on women. I’ll be speaking alongside Cindy Gallop, Jane Evans, and other incredible advocates. Register for the Deep Dive

📅 Job Board Predictions 2026 I’m also joining the Jobiqo webinar to discuss how compliance and chatbots will interact in the modern recruiting space. Grab a spot

Until next time,

Martyn

H.A.I.R. (AI in HR)

Putting the AI in HR. Safely.

Here's how H.A.I.R. can help you put the AI in HR:

1. H.A.I.R. Newsletter: get authoritative, pragmatic, and highly valuable insights on AI in HR directly to your inbox. Subscribe now.

2. AI Governance QuickScore Assessment: understand your organisation's AI governance maturity in minutes and identify key areas for improvement. Take your QuickScore here.

3. Advisory Services: implement robust AI Governance, Risk, and Compliance (GRC) with programmes designed for HR and Talent Acquisition leaders. Contact us for a consultation.

4. Measure Your Team's AI Readiness with genAssess: stop guessing and start measuring your team's practical AI application skills. Discover genAssess.