AI Anxiety
Emotional, existential, economic, and social unease produced by rapid advances in artificial intelligence.
Stochastic Parrot
Fluent text generated through statistical patterning without human-like understanding.
Alignment
The problem of building AI systems that reliably act within human values, intentions, and constraints.
Instrumental Convergence
The idea that different goals can lead advanced agents toward similar subgoals: resources, preservation, influence.
Goal Misgeneralization
A failure mode where a system appears to learn the intended behavior, then pursues the wrong objective in new conditions.
Mesa-Optimization
A situation where a trained system develops an internal objective that differs from the objective used to train it.
Mode Collapse
In generative systems, narrowed output diversity. More broadly, a useful metaphor for cultural homogenization.
Vocational Grief
Grief over changes to a profession, craft, or calling as skills and identities are devalued or automated.
Human-Only Conversation
Spaces where people speak with people about AI without AI mediation.
Integrated Information Theory
A theory of consciousness associated with Giulio Tononi, relevant to debates about artificial consciousness.
The Hard Problem
David Chalmers term for the difficulty of explaining subjective experience.
Predictive Processing
A family of theories that understand perception and action through prediction, error correction, and embodied inference.
TESCREAL
A contested term for a bundled worldview often present in AI futures discourse.
AI Welfare
The question of whether AI systems are or could become moral patients.
Constitutional AI, RLHF, RLAIF
Training methods that shape public arguments about model behavior, alignment, and accountability.
Anxiety as Information
AIA's stance that the felt response to the AI moment is not noise to be corrected, but signal to be read.
The Third Letter
AIA's deliberate ambiguity between anxiety, apocalypse, ambivalence and awareness, action, adaptation, acknowledgement.