Martin Keen

• Affiliation: ibm (ibm-technology)

• Expertise:

  • multimodal-ai, Large Language Models (LLMs)
  • AI Risk Management, Data Leakage Prevention
  • AI Agents, Agent Memory Architectures

• Key Concepts & Teaching:

  • Multimodal AI:
    • Defining modality in AI context (encompassing text, images, audio, lidar, and thermal imaging).
    • Mechanisms of how llms ingest and generate various data modalities.
  • Enterprise AI Risks:
    • Identification of five critical AI risks capable of causing career-ending consequences or organizational liability.
    • Focus on data leakage prevention as a primary concern in IBM’s AI governance framework.
  • AI Agent Memory (CoALA Framework):
    • Framework for four essential types of memory for AI agents, analogized to human cognition.
    • Discusses short-term memory, factual knowledge retrieval, and learned skills within agent architecture.
    • See: AI Agent Memory Types: CoALA Framework Overview

References

• 2026 04 10 Multimodal AI C
• 2026 05 27 AI Agent Memory Types: CoALA Framework Overview (Video: IBM Technology)