Gen X grew up in an analog world and learned digital life as it arrived. We played outside until dark, fixed what broke, waited for dial-up, then helped build the early web. Now, near 50, we are learning AI again. We remember the before, and we help shape the after. #ThaerMakdsi

AI Agent Audit Trail: 5 Fields Your Log Is Missing Five fields that turn an agent log into an audit trail, and the one rule that beats all five. Swipe. #AIGovernance #AIOversight #AIAuditTrail

An AI agent's work is gone once it's done. All you keep is the log. If it can't say what, why, and for whom, it's a text file. #AIGovernance #AIAuditTrail

One wrong answer, in an ad, to a kid's question about a telescope. The day it was reported, Alphabet lost more than $100 billion in market value. Nobody would have put that answer in the high-stakes tier. #AIGovernance #AIOversight #AIRisk

Your oversight tiers look fine on paper. Five drills to see where the scrutiny actually lands. Swipe. Thaer Makdsi, 25 years making systems trustworthy. NourFide: keeping autonomous AI under control. Light you can trust. #AIGovernance #AIOversight #HumanInTheLoop

Your reviewers scrutinize the outputs that look important. The risk sits in the ones that can promise, pay, or trigger something. Thaer Makdsi, 25 years making systems trustworthy. NourFide: keeping autonomous AI under control. Light you can trust. #AIGovernance #HumanInTheLoop

Most canceled agentic AI projects did not fail. The agent did the task. The pilot njust never collected the evidence the gate asks for. #AIGovernance #AgenticAI

Five ways an agentic AI pilot dies before it reaches production, and the move that catches each one in week two. Swipe. Thaer Makdsi, 25 years making systems trustworthy. #NourFide: keeping autonomous AI under control. Light you can trust. #AIGovernance #AIOversight #AgenticAI

Most canceled agentic AI projects did not fail. The agent did the task. The pilot just never collected the evidence the gate asks for. Thaer Makdsi, 25 years making systems trustworthy. NourFide: keeping autonomous AI under control. Light you can trust. #AIGovernance #AgenticAI

Automation was safe because it stayed what you tested. Self-learning AI never does. Your validation has an expiry date. Thaer Makdsi, 25 years making systems trustworthy. #NourFide: keeping autonomous AI under control. Light you can trust. #AIGovernance #AgenticAI #ThaerMakdsi

Your AI learns. Your validation does not. Score any learning system on five lines before you trust last quarter's test again. Swipe. #NourFide: keeping autonomous AI under control. Light you can trust. #AIGovernance #AIOversight #AgenticAI #ThaerMakdsi

Automation was safe because it stayed what you tested. Self-learning AI never does. Your validation has an expiry date. Thaer Makdsi, 25 years making systems trustworthy. #NourFide: keeping autonomous AI under control. Light you can trust. #AIGovernance #AgenticAI #ThaerMakdsi

Five AI guardrails that only watch, and the swap that makes each one stop the action. Swipe. #AIGovernance #AIOversight #AIGuardrails #ThaerMakdsi

What happens when your AI agent ignores its guardrail? If it's "we'd see it in the logs," that's not a guardrail. It's a receipt. #AIGovernance #AIGuardrails

Five AI Act duties that read like paperwork and turn out to be build work. Swipe, then sort your own list. #AIGovernance #AIOversight #EUAIAct SLIDES: cover 'Five AI Act duties / that look like paperwork'; kicker 'A sorting sheet. Save this.'

The EU gave high-risk AI 16 more months. That helps you write documents. It can't create the logs your system never kept. #AIGovernance #EUAIAct

Before you call it human oversight, run the gate through these 5. Most fail on the first one. Swipe. #AIGovernance #AIOversight #HumanInTheLoop

What would your reviewer actually have to do to catch a wrong answer? If nothing, that is a signature, not oversight. #AIGovernance #HumanInTheLoop

The bill for skipping AI governance always arrives under another name. 5 places it hides. Swipe. Thaer Makdsi, 25 years making systems trustworthy. NourFide: keeping autonomous AI under control. Light you can trust. #AIgovernance #AgenticAI

The "new" AI governance principles are older than the internet. Five rules, their prior art, and what to steal from each. Swipe. #NourFide: keeping autonomous AI under control. Light you can trust. #AIgovernance #ResponsibleAI #AIagents #SafetyEngineering #ThaerMakdsi

The "new" AI governance principles: human oversight, bounded authority, audit trails, prevention, range testing. It's not invention. It's inheritance. Use that. #AIgovernance #ResponsibleAI #AIagents #SafetyEngineering