Lardi & Partner launches AI governance framework focused on cognitive risk

Lardi & Partner Consulting GmbH has published a new framework for enterprise AI governance that targets how AI affects attention, creativity and accountability in workers. The Zurich-based firm says EEG-backed pilot data shows AI can measurably change cognitive performance, creating a gap in current oversight models.

Why it matters: - Lardi & Partner says enterprise AI governance has focused on systems, security and compliance, while overlooking how AI changes the people using it. - The company argues that cognitive effects such as reduced attention, creativity and critical thinking create a workplace risk that existing oversight frameworks do not fully address. - The framework is meant to help organizations keep AI efficiency gains without weakening human judgment and accountability.

What happened: - Lardi & Partner Consulting GmbH announced the publication of the Cognitive Governance Framework for AI Use in the Enterprise, or CGFAI. - The firm unveiled the framework in Zurich on July 8, 2026. - The company based the framework on a pilot study using wearable EEG technology to measure brain activity while professionals completed workplace tasks with and without AI assistance.

The details: - The EEG study recorded an 84.1% drop in creative neural activity when AI was introduced, and the decline was consistent across participants regardless of prior AI experience. - The study found a 12.7% decline in measured attention, with participants unaware the reduction was happening. - The research recorded a 543% surge in the familiarity response, which made AI output feel correct before it had been properly evaluated. - The study found a near-zero correlation between starting energy level and AI-assisted performance, which Lardi & Partner interprets as evidence that AI can support cognitive resilience. - CGFAI is built around five layers: guiding principles, defined roles and authority, a biometrically grounded cognitive risk register, internal controls and operational standards. - The operational standards include 10 Rules for Human-AI Cognitive Excellence, a Four Cognitive Modes framework and a 24-week implementation roadmap with actions for individuals, team leaders and senior leadership. - The framework is designed to integrate with existing AI governance programs. - Lardi & Partner says CGFAI addresses the cognitive dimension of the NIST AI RMF Govern function, supports the EU AI Act’s human oversight requirements and adds a human cognitive layer to ISO/IEC 42001. - Kamales Lardi, CEO of Lardi & Partner Consulting, said the company is focusing on what AI does to the human brain during use and called that a governance issue.

Between the lines: - The framework reframes AI governance from a primarily technical and legal discipline into one that also measures human cognition. - The EU AI Act’s requirement for a human in the loop does not necessarily guarantee meaningful human review if attention and critical thinking are already reduced. - The company is positioning cognitive performance as a measurable control point, not just a soft-management concern. - The framework’s language suggests a push to formalize AI use habits inside organizations, rather than leaving review practices to individual judgment.

What's next: - Lardi & Partner says organizations can adopt the framework alongside existing AI governance programs. - The firm is promoting the CGFAI as an implementation model for teams and leadership, not just a policy document. - The company’s next step appears to be broader adoption of the framework across enterprises using generative AI and other AI-assisted workflows.

The bottom line: - Lardi & Partner is arguing that AI governance is incomplete unless it accounts for the cognitive state of the people using the tools, not just the tools themselves.