A Review of Generative AI in Aquaculture: Foundations, Applications, and Future Directions for Smart and Sustainable Farming

Waseem Akram, Muhayy Ud Din, Lyes Saad Saoud, Irfan Hussain · 2025

Read A Review of Generative AI in Aquaculture: Foundations, Applications, and Future Directions for Smart and Sustainable Farming by Waseem Akram, Muhayy Ud Din, Lyes Saad Saoud, Irfan Hussain (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2507-11974.txt.

Introduction

Akram and co-authors review generative artificial intelligence for fish farming, published in 2025, calling it the field's first full synthesis on the subject. Generative models can fabricate text, images, audio, and simulated data. Farms are adopting them as the industry digitizes under a smart farming push. Uses span watching water conditions, steering robots, diagnosing disease, planning sites, writing reports, and reading markets.

The review lays out the core model families first: diffusion models, transformers, and retrieval systems. It then walks each application area, weighing maturity and gaps. Fish farming feeds much of the world and wastes heavily when run blind. AI that synthesizes and summarizes data could tighten that management. The authors close with future directions, from better sensors to trustworthy automation. The paper is a map for researchers entering the water.

Catalog record

CREATORWaseem Akram, Muhayy Ud Din, Lyes Saad Saoud, Irfan Hussain
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE92 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2507-11974

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

A Review of Generative AI in Aquaculture: Foundations, Applications, and Future Directions for Smart and Sustainable Farming Waseem Akram, Muhayy Ud Din, Lyes Saad Saoud, Irfan Hussain Abstract Generative Artificial Intelligence (GAI) has rapidly emerged as a transformative force in aquaculture, enabling intelligent synthesis of multimodal data, including text, images, audio, and simulation outputs for smarter, more adaptive decision-making. As the aquaculture industry shifts toward data-driven, automation and digital integration operations under the Aquaculture 4.0 paradigm, GAI models offer novel opportunities across environmental monitoring, robotics, disease diagnostics, infrastructure planning, reporting, and market analysis. This review presents the first comprehensive synthesis of GAI applications in aquaculture, encompassing foundational architectures (e.g., diffusion models, transformers, and retrieval augmented generation), experimental systems, pilot deployments, and real-world use cases. We highlight GAI’s growing role in enabling underwater perception, digital twin modeling, and autonomous planning for remotely operated vehicle (ROV) missions. We also provide an updated application taxonomy that spans sensing, control, optimization, communication, and regulatory compliance. Beyond technical capabilities, we analyze key limitations, including limited data availability, real-time performance constraints, trust and explainability, environmental costs, and regulatory uncertainty. We also identify pathways for overcoming these through multimodal modeling, federated learning, and domain-specific pretraining. Additionally, we provide an in-depth case study on the integration of GAI in marine robotics and control systems and propose future research directions for ethical, scalable, and sustainable GAI adoption. Our contributions include a structured mapping of GAI techniques to aquaculture tasks, a critical analysis of enabling technologies and barriers, and a roadmap toward responsible innovation. This review positions GAI not merely as a tool but as a critical enabler of smart, resilient, and environmentally aligned aquaculture systems. keywords Aquaculture ,Marine Robots ,Generative AI ,Autonomous systems ,Large Language Models † † credit: Conceptualization, Methodology, Software, Writing - Original Draft, Formal Analysis, Writing - Reviews and Editing † † credit: Methodology, Software - Original Draft, Investigation † † credit: Methodology, Software - Original Draft, Investigation † † credit: Methodology, Writing - Original Draft, Investigation † † corresponding: Corresponding author: I. Hussain ([email protected]) 1 Introduction The global aquaculture industry, often referred to as the “blue revolution,” has undergone rapid expansion over recent decades, becoming a critical pillar for food security, nutrition, and economic development [ 1 ] . As the demand for aquatic food continues to rise, aquaculture systems must evolve beyond traditional methods toward more intelligent, resilient, and sustainable paradigms [ 2 ] . Smart aquaculture or Aquaculture 4.0 uses Internet of Things (IoT) technologies, data analytics, robotics, and artificial intelligence (AI) to enhance productivity, mitigate environmental impact, and support real-time, data-driven decision-making [ 3 ] . Nevertheless, the sector continues to grapple with complex challenges such as disease outbreaks, biofouling, labor-intensive operations, and operational risks in dynamic and harsh marine environments [ 4 ] . Figure 1: Applications of Generative Artificial Intelligence (GAI) in smart aquaculture, spanning automation, autonomous robotics, decision support, and optimization. To overcome these limitations, the industry must shift toward more adaptive and predictive frameworks capable of synthesizing complex data streams. Generative Artificial Intelligence (GAI) offers such transformative potential [ 5 ] . By enabling reasoning across modalities, including text, imagery, audio, and simulation, GAI facilitates novel use cases ranging from synthetic data generation for fish behavior modeling [ 6 ] , disease diagnosis, and feed optimization, to real-time decision-making via natural language interfaces [ 7 , 8 ] . Foundation models such as ChatGPT, DALL·E, and diffusion-based generators are increasingly being applied as core building blocks for next-generation aquaculture platforms [ 9 , 10 ] . Figure 2: Structure of the review, organized around key thematic areas in which GAI contributes to aquaculture. The paper covers core models, applications, use cases, challenges, and future research directions. As illustrated in Figure 1 , and further detailed in Table 2 , GAI enables a wide array of capabilities across the aquaculture lifecycle including autonomous underwater inspection, multimodal sensor fusion, farm layout optimization, breeding strategy refinement, and blockchain-integrated traceability systems. GAI also plays an enabling role in robotic mission planning and fault recovery, supporting the transition toward highly autonomous and context-aware aquaculture operations [ 11 , 12 ] . Despite increasing interest in generative technologies, their systematic integration in aquaculture remains limited. Prior reviews (e.g., [ 13 ] ) primarily focus on individual applications such as feed management or fish health, leaving a broader synthesis of robotics, planning, and digital transformation unexplored. To address this gap, this review offers the first comprehensive framework connecting GAI to all key layers of smart aquaculture from perception and automation to policy-level decision support. Main contributions of this review • We provide a technical overview of core GAI models, including diffusion models, vision-language foundation architectures, and retrieval-augmented generation (RAG) systems relevant to aquaculture engineering. • We categorize and critically analyze diverse use cases where GAI addresses major aquaculture challenges, including infrastructure monitoring, water quality forecasting, breeding optimization, and stakeholder communication. • We highlight real-world prototypes and emerging platforms, such as GAI-powered ROV mission planners, traceability systems integrated with blockchain, and multilingual advisory bots. • We present a detailed tabular synthesis (Table 2 ) linking GAI models to their application domains, datasets, and evaluation benchmarks across aquaculture subsectors. • We identify technical, ethical, and infrastructural barriers to deployment, and outline a roadmap for the scalable and responsible adoption of GAI in next-generation aquaculture systems aligned with Aquaculture 4.0. The remainder of this paper is organized as follows, with its structure illustrated in Figure 2 . Section 2 provides a concise overview of GAI, while Section 3 examines its applications within the aquaculture sector. Current research and case studies are discussed in Section 4 , followed by an exploration of existing limitations in Section 5 . Future research directions are presented in Section 6 , and the paper concludes with a summary of key findings in Section 7 . 2 Overview of GAI GAI has rapidly emerged as a transformative paradigm across scientific and engineering disciplines, shifting the role of AI from analytical reasoning to autonomous generation of coherent, multimodal content. These models learn and replicate the probabilistic structure of complex datasets to synthesize diverse outputs ranging from text and images to audio, video, simulation environments, and executable code [ 14 , 15 ] . Core GAI model architectures include diffusion models, transformer-based Large Language Models (LLMs), Generative Adversarial Networks (GANs), and Variational Autoencoders (VAEs). Each offers distinct strengths: GANs excel at generating high-fidelity imagery, VAEs provide interpretable latent spaces for structured content, diffusion models support progressive denoising for photorealistic outputs, and transformers deliver scalable, context-aware reasoning across modalities [ 16 , 17 , 18 ] . The increasing relevance of GAI to aquaculture is evidenced by a dramatic rise in related scientific literature over the past decade. As shown in Figure 3 , the number of scopus-indexed publications containing the terms “GAI” or “AI” in Aquaculture, has grown from just 130 documents in 2021 to over 451 in 2024, with 256 already indexed in the first half of 2025. This upward trajectory reflects the growing convergence of intelligent automation, sustainability, and precision monitoring in the aquaculture domain [ 19 , 20 , 21 ] . This section introduces the foundational GAI model families, outlines their underlying

Read the complete text