Concerns about AI its potentially catastrophic consequences for humanity are justified. In fact, Anthropic CEO Dario Amodei has called for deliberately pacing the advancement of frontier AI so safeguards have time to catch up. His proposals include independent evaluators, stronger alignment practices, and greater industry and government coordination. However, even as these safeguards are considered, we should examine a source of great risk hiding in plain sight: the psychological makeup and motivational drivers of the teams creating and governing AI.
Read more Politics: Tom Selleck gets it right
We are investing billions to evaluate the risks embedded in AI models. Who is evaluating the inherent risks of the humans building, governing, and deciding how those models are deployed? The risk may not reside merely in individuals. The greater threat resides in homogeneous teams.
When the same behavioral predispositions are concentrated across a leadership or development team, there may be too few natural counterweights in the room. Who slows the decision? Who questions the assumptions? Who anticipates unintended consequences? Who insists on boundaries when everyone else is energized by breaking them?
The urgency is real. Although humans remain substantially in control, AI capabilities are advancing quickly. Prevention has to happen while human decisions can still meaningfully shape the trajectory. We need to examine the human guardrails now.
The very characteristics that make people exceptional at pushing boundaries may make it unwise to rely upon those same people to establish the boundaries. Technical brilliance should qualify someone to build AI. It should not automatically qualify him to determine how much risk society should accept from it.
Every roller coaster needs brakes.
For more than three decades, I have studied how inherent personality strengths, risks, motivation, and leadership team composition affect organizational outcomes. AI has not changed those human dynamics. It has dramatically increased the consequences of getting them wrong.
This isn’t an argument for replacing brilliant innovators with cautious people. It is an argument for balance. When nearly everyone at the table is highly inquisitive, adventurous, willing to break rules, comfortable taking significant risks, and energized by power and competition, adding one cautious voice isn’t enough. That person is likely to be shut down or ignored by the prevailing culture.
We need multiple credible counterbalances with the authority to influence decisions — and a leadership team aligned around a mission that defines not only what it hopes to accomplish, but what risks it is unwilling to take. That requires assessing the collective human risk of leadership and R&D teams and deliberately building counterbalancing capabilities where they are lacking.
Read more Egypt’s Luxury Property Builders Have Some Questionable Connections
Conventional methods of evaluating executives rarely uncover these risks. Résumés, interviews, performance histories, and background checks provide basic information about experience and observable behavior but offer limited insight into the inherent behavioral tendencies, talent gaps, or imbalances that can put an organization at risk. We need scientifically valid measures that assess inherent strengths and risks, individually and collectively, before those risks become consequences.
More than two decades ago, I made a similar argument about corporate risk. In an article for Risk Management Magazine, I compared assessment data from former Enron executives with those of felons incarcerated in a maximum-security prison and found their profiles remarkably similar. The lesson was that organizations were evaluating financial, operational, and market risks while largely ignoring the measurable behavioral risks sitting inside the executive team.
Recently, I facilitated a development session with the C-suite of a Silicon Valley AI company. The assessment data revealed an extraordinarily homogeneous team: brilliant innovators and risk-takers, but with little counterbalancing operational capability. The executives recognized it immediately. Their very strengths — the qualities that fueled innovation — were also contributing to execution problems and putting the business at risk.
Their solution wasn’t to constrain the innovators. It was to bring more operational, practical, results-oriented talent from the level below into the leadership equation. They needed balance, not less brilliance.
The tools already exist to assess AI executive and R&D teams, identify behavioral strengths and risks, and determine whether teams have the counterbalances needed for sound decision-making. The question is whether we will use them.
AI presents the same human problem at a new scale and speed. The principles are not new. The stakes are. If we want AI to serve humanity well, we must devote as much attention to the judgment, balance, and boundaries of the people leading the technology as we do to the technology itself.
Read more No Body. No Jury. And the Doctor Who Thought He Was Smarter Than Murder
Image via Pexels.