# What AIs are not Learning (and Why)

arXiv: https://arxiv.org/abs/2404.04267 · HTML: https://mlsanity.com/paper/2404.04267/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 8 of its 75 references that are in the map.
- First posted on arXiv 19 Mar 2024.

## Builds on (most cited first, 8 of 8)

- [Scaling Laws for Neural Language Models](https://mlsanity.com/paper/2001.08361/) (2020)
- [Language Models as Zero-Shot Planners: Extracting Actionable Knowledge for Embodied Agents](https://mlsanity.com/paper/2201.07207/) (2022)
- [Socratic Models: Composing Zero-Shot Multimodal Reasoning with Language](https://mlsanity.com/paper/2204.00598/) (2022)
- [Foundation Models in Robotics: Applications, Challenges, and the Future](https://mlsanity.com/paper/2312.07843/) (2023)
- [Why AI is Harder Than We Think](https://mlsanity.com/paper/2104.12871/) (2021)
- [Adaptive Robot-Assisted Feeding: An Online Learning Framework for Acquiring Previously Unseen Food Items](https://mlsanity.com/paper/1908.07088/) (2019)
- [Bodily aware soft robots: integration of proprioceptive and exteroceptive sensors](https://mlsanity.com/paper/1710.05419/) (2017)
- [Bootstrapping Developmental AIs: From Simple Competences to Intelligent Human-Compatible AIs](https://mlsanity.com/paper/2308.04586/) (2023)

## Abstract (arXiv)

Today's robots do not learn the general skills needed for such services as providing home care, being nursing assistants, or doing household chores. Addressing such aspirational goals requires improving how AIs and robots are created. Today's mainstream AIs are not created by agents learning from experiences doing real world tasks and interacting with people. They do not learn by sensing, acting, doing experiments, and collaborating. This paper investigates what aspirational service robots will need to know. It recommends developing experiential (robotic) foundation models (FMs) for bootstrapping them.