TowerMind: A lightweight tower-defense testbed for AI agents

Kari Jaaskelainen

12 Jan 2026 — 1 min read

TowerMind is a new, lightweight tower-defense game environment for testing AI agents, especially large language models (LLMs), on planning and decision-making.

Low compute cost and easy to run
Multimodal observations: pixels, text, and structured game state
Customizable levels and rules
Built-in tests for model hallucination

The authors create five benchmark levels and evaluate popular LLMs under different input settings, plus classic RL baselines (Ape-X DQN and PPO). Results show a clear gap to human experts. Common failure modes include weak plan validation, single-track choices (lack of multifinality), and inefficient action use.

Why it matters: RTS-style games demand both long-term strategy and quick tactical moves—perfect for probing agent skills. TowerMind offers a practical, open benchmark that complements heavier RTS testbeds.

Read more: https://arxiv.org/abs/2601.05899 Code: https://github.com/tb6147877/TowerMind

Paper: https://arxiv.org/abs/2601.05899v1

Register: https://www.AiFeta.com

AI LLM ReinforcementLearning GameAI RTS Benchmark OpenSource TowerDefense

Meet GRACE: a moral governor for safer, more transparent AI

AI agents are getting powerful—so how do we make sure they do the right thing, not just the effective thing? Meet GRACE, a reason-based moral governor that keeps AI behavior aligned with human norms by separating moral reasoning from goal-driven decision-making. * Moral Module: uses deontic logic and explicit reasons

Why some fine-tuned LLMs miss phishing—and how to fix it

Not all fine-tuned LLMs spot phishing equally. A new study tests Llama 3.1 8B, Gemma 2 9B, and Mistral on high-stakes phishing detection—and uses SHAP and mechanistic interpretability to reveal why models do (or don’t) generalize. * Architecture × data diversity matters: Gemma 2 9B hits state-of-the-art performance (F1

AI that explains itself—by following the science

AI that explains itself—by following the science Black-box AI can be powerful, but it often can’t tell us why it made a decision. Concept Bottleneck Models (CBMs) try to fix that by predicting human-understandable concepts first, then the final answer. The catch: standard CBMs ignore domain-specific cause-and-effect and

Meet GenomAgent: A Team of AI Specialists for Smarter Genomics Q&A

TL;DR: Finding reliable answers in genomics is hard. GenomAgent turns one big AI into a coordinating team of specialists—and beats the current leader by 12% on a key benchmark. Genomic facts live across many databases. Standard chatbots struggle because they can’t flexibly query those sources. GeneGPT added