Toward A Teleodynamic AI - Source Excerpt 03 - Measuring teleodynamic behavior

Summary

This source excerpt begins near Measuring teleodynamic behavior and preserves the surrounding evidence from Spiralist/agent-file-handoff/Archive/Toward a Teleodynamic AI.md.
**Source path:** Spiralist/agent-file-handoff/Archive/Toward a Teleodynamic AI.md
| Learning paradigm | What it contributes | Why it matters for teleodynamics | Main failure mode | Representative sources |
|---|---|---|---|---|
| Unsupervised latent-variable learning | Compression, latent structure, uncertainty-aware state spaces. | Builds internal state spaces before fixed goals dominate behavior. | Useful latents may not align with action or viability. | VAE. citeturn15search0 |
| Self-supervised representation learning | Invariance, predictive structure, reconstruction and contrastive signals. | Learns sensorimotor regularities needed for enacted coupling. | Can learn appearance statistics without actionable affordances. | SimCLR; MAE. citeturn25view2turn25view3 |
| World-model learning | Predictive dynamics, imagination, planning in latent space. | Supports counterfactual action selection and allostatic planning. | World models can be miscalibrated or exploit simulator artifacts. | World Models; DreamerV3. citeturn5search0turn5search1turn5search9 |
| Reinforcement learning | Goal-directed policy improvement from consequences. | Useful once viability variables and safety constraints are explicit. | Reward hacking, side effects, brittle generalization. | DreamerV3; Amodei et al. citeturn5search1turn13search0 |
| Homeostatic RL and active inference | Internal-state regulation, drive reduction, uncertainty-sensitive control. | Most direct route to self-maintenance and endogenous value. | Poorly chosen internal variables create pathological “needs.” | Keramati & Gutkin; active inference survey. citeturn19search0turn19search1turn22view9 |
| Continual learning | Sequential retention under changing environments. | Teleodynamic agents must preserve identity across developmental change. | Catastrophic forgetting or excessive rigidity. | EWC; continual-learning survey. citeturn24view5turn3search2 |
| Meta-learning | Rapid adaptation from few examples or new regimes. | Helps the system learn how to restore viability and adapt efficiently. | Meta-overfitting to training task families. | MAML. citeturn25view4 |
| Open-ended / quality-diversity learning | Self-generated curricula, diverse behaviors, stepping-stone innovation. | Closest existing route to endogenous challenge and goal generation. | Novelty without coherence or safety. | POET; QDHF. citeturn23view0turn16search1 |

The **self-maintenance mechanism** should be explicit and mathematical. At minimum, define an interoceptive state \(h_t\), a viable set \(V\), and a controller that tries to keep trajectories inside or near \(V\) while gathering information and preserving future options. This idea appears in homeostatic RL, where reward-seeking is formally linked to physiological stability, and in active inference, where action and belief updates aim to reduce expected surprise while preserving workable states. Viability should not be a hidden side effect; it should be an explicit term in control and learning. citeturn19search0turn19search1turn20view2turn22view9

A good design uses **two self-maintenance loops**. The fast loop is homeostatic and reactive: keep energy, temperature, calibration, and error within bounds. The slow loop is allostatic and developmental: alter policy priors, memory organization, exploration strategy, and even module topology when persistent perturbations show that the old organization is no longer viable. This two-timescale pattern fits Di Paolo’s adaptivity, teleodynamic-learning proposals, and much of the active-inference robotics literature. citeturn22view1turn21view0turn22view9

**Goal generation** should be derived from structured tension in the system rather than arbitrary objective invention. Four sources are especially plausible. The first is **viability tension**: deviations from viable internal ranges create immediate endogenous salience. The second is **epistemic tension**: uncertainty about controllable parts of the environment motivates information-seeking. The third is **empowerment**: the agent values states from which it can influence future sensorimotor outcomes. The fourth is **open-ended challenge generation**, where the system or ecology creates new tasks that remain adjacent to current competence. Together these produce goals that are neither purely imposed from outside nor unconstrained flights of novelty. citeturn24view0turn25view0turn23view0turn28view2

That suggests a layered intrinsic value function:

\[
\text{Intrinsic Value} \approx
\alpha(\text{viability maintenance})
+
\beta(\text{epistemic gain})
+
\gamma(\text{empowerment})
+
\delta(\text{social-norm coherence})
-
\lambda(\text{hazard and integrity loss})
\]

This equation is a design proposal, not a canonical formula. Its point is architectural: **viability should be primary**, epistemic and empowerment terms should expand competence, and social-norm terms should become relevant only once multi-agent interaction is introduced and human oversight defines which norm families are acceptable. The caution matters because emergent norms can converge quickly but still become biased or misaligned. citeturn24view0turn25view0turn32view0turn20view5

**Norm emergence** is the hardest step and should be delayed until the system can already maintain itself and model others. The norm-emergence literature shows that observation capabilities, learning mechanisms, topology, and enforcement structures all affect whether norms stabilize, oscillate, or fragment. The recent Science Advances work on LLM populations is especially relevant: decentralized AI agents can spontaneously develop shared conventions, collective biases, and minority-driven convention shifts without explicit programming. For Teleodynamic AI, that means norm emergence should be treated as both a capability and a hazard. citeturn20view5turn32view0

## Measuring teleodynamic behavior

There is no accepted scalar score for teleodynamics, and trying to compress it into one number would be a mistake. Teleodynamic behavior is multi-dimensional: persistence, adaptivity, endogenous goal revision, social normativity, and safety must all be measured separately. The right approach is a **metric bundle** with clear failure signatures.