Toward Teachable Autotelic Agents

Autonomous discovery and direct instruction are two distinct sources of learning in children but education sciences demonstrate that mixed approaches such as assisted discovery or guided play result in improved skill acquisition. In the field of artificial intelligence (AI), these extremes, respectively, map to autonomous agents learning from their own signals and interactive learning agents fully taught by their teachers. In between should stand teachable autonomous agents (TAA): agents that learn from both internal and teaching signals to benefit from the higher efficiency of assisted discovery. Designing such agents will enable real-world nonexpert users to orient the learning trajectories of agents toward their expectations. More fundamentally, this may also be a key step to build agents with human-level intelligence. This article presents a roadmap toward the design of teachable autonomous agents from a developmental AI perspective. Building on developmental psychology and education sciences, we start by identifying key features enabling assisted discovery processes in child-tutor interactions. This leads to the production of a checklist of features that future TAAs will need to demonstrate. The checklist allows us to precisely pinpoint the various limitations of current reinforcement learning agents and to identify the promising first steps toward TAAs. It also shows the way forward by highlighting key research directions toward the design or autonomous agents that can be taught by ordinary people via natural pedagogy.