Understanding and Predicting the Ocean using Artificial Intelligence (Second Edition)

The data foundations and governance conditions required for responsible and scalable AI in ocean science.

Topics we welcome include:
• Preparing, harmonizing, and documenting ocean data for AI training and reuse
• Implementing FAIR and CARE principles in practice — metadata, licensing, stewardship
• Indigenous data governance approaches and co-developed frameworks
• Interoperability across formats, vocabularies, platforms, and infrastructures
• LLM and agentic approaches to metadata generation and natural-language data discovery
• Governance tools or policies that enable trustworthy AI across institutions

Relevant contributors may include: Data stewards, infrastructure leads, governance and ethics practitioners, Indigenous partners, standards and policy specialists.

AI methods and model development, with an emphasis on approaches moving toward operational or applied use.

Topics we welcome include:
• AI models, hybrid physics-AI models, and emulators for ocean prediction
• Computer vision or acoustic classifiers for applications such as biodiversity monitoring, acoustic characterization and vessel detection
• Machine-learning approaches for anomaly detection or data quality assessment
• Generative methods (data augmentation, gap-filling, downscaling) and foundation models
• Explainable or interpretable AI methods applied to ocean data
• Lessons learned from moving models beyond proof-of-concept

Relevant contributors may include: Ocean modelers, AI/ML researchers, applied scientists, method developers.

Deployment, pipelines, tooling, and repositories — what is required to scale and sustain AI systems in real-world ocean contexts.

Topics we welcome include:
• Deployment of AI models into production or pre-operational environments
• Data and model pipelines, MLOps practices, and automation workflows (including agentic and LLM-based approaches)
• Infrastructure, shared tools, platforms, or repositories supporting AI use
• Integration of AI into existing operational systems or decision processes
• Lessons learned from maintaining, updating, or scaling AI services

Relevant contributors may include: Platform developers, operational teams, research software developers, infrastructure providers.

Benchmarking, explainability, risk, and use of AI in regulatory or decision-making settings.

Topics we welcome include:
• Benchmarking and evaluation frameworks for AI models
• Validation under extreme events or non-stationary conditions
• Explainability and transparency for operational or regulatory use
• Risk assessment, assurance, and quality control approaches
• Bridging AI outputs with peer-reviewed or regulatory science processes

Relevant contributors may include: Validation experts, QA/QC leads, regulatory scientists, risk and governance specialists.