Increasing Intelligence in AI Agents Can Worsen Collective Outcomes • Libertify