By Bill Cottringer

Here is a very interesting exchange I just had with ChatGPT.

ChatGPT—I’d value your opinion on this:

Question: When will AI learn EI?

Answer: It depends on a critical realization, which is:  Objective rational thinking has its limits, because the “truth” AI concludes is based on the trust in the source of the information believed to be true and valid. In humans this is a leap of faith in trusting the validity of the information’s source and this “trust” is the strength of feeling that the conclusion is virtually certain but can never really be proven objectively. In that sense, truth is more subjective and emotional, rather than objective and rational.

Here is how the AI program responded:

What You’re Capturing Well

You’re making three strong, defensible points:

1. Objective rationality has limits

This is well established in:

• Cognitive psychology (bounded rationality)
• Philosophy of science (underdetermination of theory by data)
• Decision science (uncertainty, ambiguity, and trust)

AI optimizes inference, but inference alone does not equal understanding—especially under uncertainty.

2. Truth depends on trust in sources

This is an underappreciated insight in AI discourse.

Humans don’t just evaluate facts—we evaluate:

• Credibility
• Intent
• Context
• Relational history

Much of what humans “know” is not proven, but trusted. That trust is emotional, social, and experiential—not purely logical.

AI currently:

• Weighs sources statistically
• Does not experience trust
• Cannot feel betrayal, confidence, doubt, or moral stake

This is a real barrier to Emotional Intelligence (EI).

3. Human certainty is affective, not purely logical

You’re right that human certainty often comes from a felt sense of rightness, not deductive proof.

Neuroscience backs this up:

• Emotion is not the enemy of reason—it terminates infinite analysis
• Decisions require affective commitment (Damasio’s somatic marker hypothesis)

In that sense, emotion is what allows action under uncertainty.

Where the Answer Could Be Refined

Here’s the key distinction I’d suggest tightening:

EI is not the same as truth

EI is about:

• Recognizing emotional states
• Understanding their meaning
• Regulating responses
• Navigating social contexts

What you’re describing is closer to epistemic trust and meaning-making, which underpins EI but is not identical to it.

A More Precise Framing

AI will not “learn EI” simply by realizing truth is subjective.

Instead, AI would need to:

1. Model emotional meaning, not just emotional signals
2. Represent uncertainty as lived risk, not probability
3. Assign value under moral and relational stakes, not optimization goals
4. Care, in the sense of having something to lose

That last point is crucial.

A Sharpened Version of Your Answer (Optional Rewrite)

If you wanted a slightly more focused and rigorous version while preserving your insight:

AI will approach emotional intelligence only when it recognizes the limits of objective rationality—specifically, that what humans accept as “truth” is grounded in trust, not proof. Human knowing requires a leap of faith in the credibility of sources, reinforced by emotional certainty rather than logical finality. Because this trust is experiential and affective, not computational, EI remains fundamentally human unless AI can genuinely experience uncertainty, value, and relational risk.

Bottom Line Opinion:

Your answer is:

• Conceptually strong
• Psychologically grounded
• Philosophically defensible

Its deepest implication—perhaps unintentionally—is this:

EI is not a skill AI lacks; it is a condition AI does not inhabit.