Michael Paik
Founder, GRC Solutions Korea
Attorney, AI Auditor & Architect
About
GRC Solutions Korea bridges algorithmic audit, legal structuring and operational execution to build enterprise AI Governance, Risk Management & Compliance (GRC) systems for defensible control under scrutiny.
5 Projects
AI GRC Copilot
The AI GRC Copilot helps generate a stack of 30 basic artifacts for use in AI GRC. Not everyone needs every item. If you are running high-risk AI in a grown-up organization, this is the map. Governance is, among other things, a documentation engine that produces decisions and proof. If you cannot point to the artifact, the owner, the control and the log, you do not have governance. You have hope. The prompt will not do the job for you. You supply facts, constraints, decisions and approvals. The model drafts and stress-tests. You own the risk. Delegation is not abdication. ๐๐ถ๐ป๐ฎ๐น ๐๐ถ๐ฎ๐ฏ๐ถ๐น๐ถ๐๐ ๐ฟ๐ฒ๐๐๐ ๐๐ถ๐๐ต ๐๐ต๐ฒ ๐๐๐บ๐ฎ๐ป.
RedCap-00
RedCap-00 is the first tool in ๐๐ฅ๐ ๐ป๐ฒ๐ ๐โข: a free public self-check, evidence-capped and conservative by design. Use it to evaluate whether your organization can make real operational choices inside 72 hours under disruption. It focuses on five moves: safe pause, treasury reroute, data switchboard, substitution of critical dependencies, and clean market exit.
Contract-Mechanism-Review-Assistant
This is the spellbook for the Contract Mechanism Review Assistant. It reads slowly, thinks conservatively, and writes with discipline. It treats every agreement as: a mechanism for risk transfer a business plan in clauses a memorialization of future performance It produces clause-by-clause issue spotting, insert-ready edits, fallback positions, and a short deal-risk summary you can paste into an email.
origami-method
๐ข๐ฟ๐ถ๐ด๐ฎ๐บ๐ถ ๐ ๐ฒ๐๐ต๐ผ๐ฑ Lock a one sentence target Build a context packet: objective, boundaries, definition of done Classify risk high enough to justify strict gates Define creases: inputs, outputs, exclusions Run the workflow: read โ analyze โ draft โ gatekeep, test on real cases, log minimal folds, then lock The Origami Workflow Guide helps design the dedicated prompts you need for repeatable workflows, without improvisation.
GRCnext-Copilot
GRCnextโข Copilot scores evidence-backed โGlobal Optionalityโ across critical services (Pipes, Switches, Exits) and outputs dossiers, exit tactics, playbooks, drills, and a 90-day backlog.
Media Appearances
Manufacturing Knowledge, Governance and Security Thought Leadership Series
Gravitate AI
Interview on preserving retiring manufacturing expertise with AI
As AI Goes Global, Corporations Turn Inward for a Unified GRC Strategy
AI Data Press
Son-U Michael Paik, AI governance architect and NY attorney, explains why a unified framework and human expertise are essential for managing global AI risk.
How In-House Counsel Can Thrive with AI
The Business Simplicity Podcast
About how legal teams can use AI to strengthen operations, reduce overwhelm, and step confidently into the future. From risk frameworks to productivity tools to mental models for AI interaction, this conversation is a practical roadmap for legal leaders and rising professionals alike.
Lessons Learned from South Korea
Dentons Legal Operations Insights
As law department leaders around the world assess how their organizations and law departments will not only recover, but thrive, in the rapidly evolving environment of this pandemic, looking at how others who were on the front line earlier in reopening their economies and offices can help prepare you as we enter this new dynamic. To share insight on this, we connect with Michael Paik, Executive Vice President, Legal Trade and Risk Management at SeAH Holdings Corp, a global steel manufacturer, based in South Korea as he shares the lessons they learned as a law department continuing to effectively support their business through this constant state of accelerating change.
Philosophy
"Human-made, for Easy Work, Simply Done."
My approach to AI rests on three principles: Informed Intent, Slow AI and Final Liability. Informed Intent defines what we are trying to accomplish, under what constraints and to what standard. Slow AI makes the process visible, deliberate and auditable, preserving appropriate human judgment and control. Final Liability recognizes that responsibility cannot be delegated to a machine: every material outcome must ultimately belong to a named human with the authority, information and ability to intervene. The practical goal is โHuman-made, for Easy Work, Simply Done.โ AI should reduce friction and complexity without obscuring authorship, judgment or accountability. The best systems make difficult work easier while keeping human purpose and responsibility intact. AI provides capability. Humans provide intent, judgment and ownership.