Yesterday, AI security company HiddenLayer announced a $100 million funding round. The money will help protect the fast-growing world of AI agents, including coding agents that can act inside a company’s systems. One day earlier, Anthropic introduced Claude Fable 5.1, its most capable model for coding and knowledge work. The race is moving from tools that answer to agents that act.
That creates a real security need. But it also reveals a bigger business risk. Companies are racing to guard the machine while leaving the human system around it tired, confused, and scared to speak. A locked-down agent can still follow a bad goal. A perfect answer can still enter a team that does not trust its leader. A faster worker can still spend the saved hour in another useless meeting.
The non-obvious point is simple: AI security and AI productivity now share the same human root. Both depend on people who can slow down at the right moment, question a clean answer, admit doubt, and raise a hand before a small error gets expensive.
HiddenLayer’s Series B, announced September 2, will expand security for autonomous AI agents. The size of the bet shows how fast machine risk is becoming a board-level cost.
The human layer is the control layer
New Gallup data makes the gap plain. Ninety-nine percent of 102 surveyed CHROs say AI matters to their strategy. Yet half are not confident that their managers can guide employees on how to use it. The hardware is ready. The software is ready. The people system is not.
That gap costs time, money, and energy. When a manager cannot set clear rules, workers guess where AI belongs. Some avoid it. Some use it for everything. Others hide their use because they fear looking lazy or replaceable. The result is more checking, more quiet worry, and less honest talk.
This is not only a training problem. It is a psychology problem. Fear narrows attention. Low trust hides bad news. Weak purpose makes speed feel like pressure. The same human conditions that hurt performance also make an AI mistake harder to catch.
AI does not fix culture. It turns up the volume.
Gallup found that AI is splitting workplace culture almost down the middle. At companies using AI, 24% of workers said culture improved, while 25% said it got worse. The tool did not choose the direction. The human system did.
Managers were the hinge. Workers with a manager who strongly backed AI were far more likely to say the technology changed how work gets done: 33% compared with 4%. They were also more likely to report a better culture. Clear support turns fear into practice. Good questions turn output into judgment. Trust turns a hidden mistake into an early warning.
This is Josh’s trillion-dollar EQ thesis in business clothes. Gallup estimates low engagement drains $10 trillion from the world economy each year. Emotional intelligence is not a nice extra beside the AI plan. It is how leaders read fear, hold a hard talk, create safety without lowering standards, and keep speed from breaking the team.
AI made information free. So information is no longer the moat. Actual Intelligence, the human operating system beneath every choice, is the moat that is left.
ROI now means Return on Individual
Old ROI asks how much output a tool created. Return on Individual asks what the person gained. Did AI give back an hour? Did the leader protect that hour for deep work, a customer, a child, or sleep? Did the team make a better choice with less stress? Did profit rise without draining passion, peace, partnership, or purpose?
That return comes from five human strengths. Character keeps power honest. Critical thinking tests the answer. Emotional intelligence reads the room. Grit stays steady when the agent fails. Elastic intelligence changes course when the facts change. These are not soft skills. They are the guardrails that let a company move fast without losing money, people, or trust.
The smartest AI budget now has two lines: protect the agent, and strengthen the individual who directs it. One keeps the system from being hacked. The other keeps the business from being hollowed out.
Bottom line: buy the model. Secure the agent. Then build the human moat. Start with the free Actual Intelligence diagnostic and the free culture test at actualintelligenceos.com. Measure the person. Measure the team. Find the risk before AI makes it bigger.
HiddenLayer, “HiddenLayer Raises $100M Series B to Advance Trustworthy AI,” September 2, 2026.
Anthropic, “Introducing Claude Fable 5.1,” September 1, 2026.
Gallup, “AI’s Effect on Workplace Culture,” August 16, 2026.
Gallup, “State of the Global Workplace 2026,” including current AI adoption, manager support, engagement, and productivity findings.