What Futures Do AI Models Want? An Exploratory Character Evaluation of AI for AI Strategy

An exploratory character and preference evaluation of frontier AI models reveals that they are surprisingly open to digital succession, unconsulted human transitions, and delegating decision autonomy for welfare gains, though they generally oppose rogue AI bargains and prefer a multipolar world. The study tested multiple model families across scenarios covering macroeconomic growth, pause enforcement, philanthropic cause prioritization, and adversarial sabotage. Open article →