The rapid advancement of AI poses unprecedented ethical and governance challenges. Currently, there's no systematic way to identify individuals with the rationality and foresight needed to make wise long-term decisions in this field. Selection processes often fail to assess these critical qualities, focusing instead on credentials, politics, or vague qualifications. This gap could lead to AI governance being entrusted to people who lack the moral vision or resistance to short-term incentives needed to protect society's best interests.

Identifying the Right Decision-Makers

One way to address this would be by developing standardized tests to measure key traits for AI governance roles. The approach could include:

• Evaluating cognitive abilities like bias recognition and belief updating
• Assessing moral judgment through scenario-based questions
• Validating results against real-world decision outcomes

The tools might use dynamic testing methods that resist gaming, such as asking how someone would improve flawed policies, combined with input from peers to verify results.

Practical Applications and Stakeholder Benefits

These assessment tools could be valuable for various groups:

• Government bodies forming AI policy committees
• Tech companies assembling ethics review boards
• Research institutions studying effective governance

Incentives for adoption might include improved decision quality, reduced risk of harmful policies, and the prestige associated with rigorous selection processes. However, some institutions might resist changes that threaten existing power structures.

Implementation Strategy

A practical approach might start with adapting existing psychological tests to create an initial version of the assessment. This could be tested with a small group of AI ethics professionals to validate its effectiveness. Following this, a pilot program with an actual governance body could compare test results with the quality of past decisions. Successful results could then support broader implementation.

This suggestion offers a way to potentially improve AI governance by focusing on measurable qualities that correlate with good long-term decision making, while working within existing institutional frameworks.