Sentiment alone misreads AI coverage — a model that is “superhuman” thrills tech blogs and alarms
safety researchers in the same breath. Instead, the Narrative Lens reads the story being told. A broad
slice of the last 30 days of English AI news is classified along four independent axes — myth, agency, time
horizon, and the capability/control balance — to reveal which narrative is winning the news cycle.
Reading the current news cycle…
Classifying recent AI headlines across four narrative axes…
1 · The Archetypal Index The Mythos
Which classic archetype is dominating coverage — a tool, a threat, an oracle, or a companion? Dominance is the 30-day share among framed stories.
2 · The Agency Attribution Index
Does the headline cast the AI as the actor (“ChatGPT passes the bar”) or as the object being acted upon (“OpenAI updates model”)? A rising machine-agency share marks a psychological shift toward treating AI as an autonomous actor.
3 · The Temporal Horizon Score
Is the conversation fixed on tomorrow’s deepfakes or next decade’s AGI? The horizon mix indexes practical utility against long-range alignment anxiety.
4 · The Capability vs. Control Matrix
Every story plotted by how powerful (Y) and how steerable (X) it frames the AI. High capability + low control is the existential-risk quadrant; high capability + high control is the tech-optimist utopia.
Framing, not fact. Each axis describes how a story is told, not whether its claims are true. The capability and control scores read the rhetoric of the headline, not a model’s measured performance.
Model-assisted classification. Headlines are labelled by a language model over a rolling 30-day English sample, so the indicators are a directional read of the news cycle rather than a census.