Numerous studies highlight that investors and other stakeholders consider animal welfare a crucial aspect of corporate social responsibility (CSR). This emphasis stems from the growing recognition of companies’ moral and ethical obligation to minimize harm to animals, particularly when their operations have a direct impact. Beyond ethical concerns, animals and other living beings play a vital role in maintaining the ecosystems that support business operations. As a result, addressing animal welfare has become an integral part of a company’s commitment to social responsibility and accountability.

The advancement of AI introduces new dimensions to CSR in relation to animal welfare. The following examples illustrate how AI can impact animal welfare and underscore the importance of integrating AI considerations into the CSR policies and practices of relevant industries.

Toxicology

Many studies indicate that AI holds significant potential to enhance toxicity risk assessment and reduce reliance on animal testing. AI models can rapidly process large and complex datasets, identifying patterns and relationships that may be challenging for humans to discern. According to Hartung (2023), AI’s data analysis capabilities can lead to more accurate toxicity predictions for new compounds, supporting precise risk assessments and potentially diminishing the necessity for animal testing.

Fjodorova et al. (2023) exemplify the application of AI techniques for toxicity assessment in real-world scenarios. The study investigates how cheminformatics and machine learning methods can evaluate the aquatic toxicity profiles of fullerene derivatives, which are utilized in nanomaterials and pharmaceuticals. Specifically, the researchers employed Counter-Propagation Artificial Neural Network (CPANN) models—a type of quantitative structure-activity relationship (QSAR) model—to predict the binding affinity of fullerene derivatives to proteins, a critical factor in assessing toxicity. The CPANN models effectively captured the relationships between the structural features of fullerene derivatives and their binding affinities to proteins, thereby validating their utility in predictive toxicology.

Similarly, Igarashi et al. (2024) developed a Graph Neural Network (GNN)-based artificial intelligence model to predict mitochondrial toxicity. GNNs offer several advantages in toxicity prediction: they learn directly from the graph representations of chemical structures, eliminating biases associated with traditional molecular descriptors; they facilitate the visualization of structural alerts through methodologies such as integrated gradients, thereby providing valuable insights into the mechanisms of toxicity by identifying key substructures associated with toxic effects; and, when combined with techniques such as bagging, GNNs effectively handle imbalanced datasets, enhancing predictive performance. This approach resulted in a high F1 score of 0.839, illustrating its effectiveness in accurately predicting mitochondrial toxicity in drug candidates.

For companies in the chemical industry and other sectors managing toxicity risks, evaluating AI technologies within their CSR framework is becoming increasingly important. By assessing how AI can enhance existing processes, businesses can strengthen their commitment to both animal welfare and public health. This not only aligns with ethical obligations but also meets growing public and stakeholder expectations for humane and sustainable practices.

Animal Husbandry

According to Rosati (2024), AI can be applied in animal husbandry in several impactful ways. For example, it can facilitate real-time monitoring by utilizing data from IoT sensors to track animal health, living conditions, and environmental factors in real time. This enables the early detection of health issues, allowing for timely interventions to improve animal welfare. Real-time monitoring also helps address unfavorable conditions such as overcrowding or unsanitary environments. Additionally, AI can optimize feeding by analyzing individual animal needs and adjusting feeding regimens accordingly, which not only improves animal welfare but also reduces waste and enhances feed efficiency, leading to better resource management.

While AI presents significant opportunities to enhance animal welfare and operational efficiency in animal husbandry, its implementation also introduces challenges that CSR initiatives must address. For example, Ajibola et al. (2024) notes that applying AI to enhance on-farm poultry welfare is not as straightforward as it may seem at this stage. AI systems are usually trained and tested in controlled research settings, but poultry farms vary significantly in structure, barn design, equipment, and animal density. These differences can complicate the deployment of AI systems, as algorithms must be specifically trained to accommodate the unique characteristics of each farm environment. Additionally, while AI algorithms typically focus on specific aspects of poultry health or behavior, ensuring animal welfare requires a holistic assessment that considers their physical health and functioning, mental well-being, and broader living conditions and social dynamics (Ajibola et al., 2024, Mellor and Beausoleil, 2015).

Biotechnology and Pharmaceuticals

AI also has the potential to reduce animal testing in the pharmaceutical and biotechnology sectors. For example, Rudroff (2024) discusses several advanced AI-based methods with the potential to replace animal experiments in neurology: organoid intelligence, which integrates lab-grown brain organoids—three-dimensional cultures of human brain cells—with AI algorithms and brain-machine interface technologies, enabling researchers to investigate disease mechanisms and evaluate treatments in systems that closely mimic human physiology; AI-enhanced computational models of neural circuits, which simulate complex brain functions, allowing the study of neurological activity and disease mechanisms without relying on animal models; and machine learning techniques that analyze large datasets to identify drug targets, predict neurotoxicity, and optimize drug candidates, significantly reducing the need for animal testing. Serrano et al. (2024) also discuss AI’s role in drug discovery and delivery, highlighting its potential to revolutionize personalized medicine. AI enables tailored treatments based on individual genetic, lifestyle, and disease characteristics. Through advanced algorithms, AI predicts patient responses to drugs and continuously optimizes dosing regimens in real-time, enhancing effectiveness and minimizing adverse effects.

Potential Harms of AI to Animal Welfare

CSR initiatives must consider the potential harms that AI technologies may pose to animal welfare. These harms can be direct or indirect. Coghlan and Parker (2023) present a comprehensive analysis of the various types of harm that AI can inflict on nonhuman animals, organized within a five-part harms framework. The following image illustrates the different types of harm, along with simplified examples for each category.

Incorporating animal welfare into AI ethics frameworks is a logical progression, given AI’s significant impact on non-human animals. Singer and Tse (2023) advocate for expanding AI ethics to encompass all sentient beings, emphasizing the moral importance of animals and the vast numbers affected by AI technologies. They propose establishing a dedicated field of research to examine AI’s impact on animals, analyze ethical implications, and explore strategies to minimize harm. This includes conducting case studies to highlight how AI influences animal welfare and identifying key ethical concerns.

In the context of livestock farming, Neethirajan (2024) proposes a human-centric approach to animal welfare. This approach emphasizes integrating AI and sensor technologies with farmers’ intrinsic knowledge to enhance animal well-being. By integrating continuous feedback from sensor data, farmers can refine care strategies that adapt to the evolving needs of livestock. Utilizing AI-driven tools to monitor health, behavior, and physiological metrics enables informed decision-making, improving care strategies. This collaborative model positions technology as an enhancer of human expertise, aiming to create a harmonious balance between technological advancements and ethical farming practices.

These perspectives underscore the necessity of broadening AI ethics to include animal welfare, promoting research and practices that protect all sentient beings affected by AI systems. CSR policies grounded in these considerations can guide organizations in developing and implementing AI technologies that enhance animal welfare, ensuring technological advancement aligns with ethical principles and societal values.

References:

Ajibola, G., Kilders, V., & Erasmus, M. A. (2024). A peep into the future: artificial intelligence for on-farm poultry welfare monitoring. Animal Frontiers, 14(6), 72–75. https://doi.org/10.1093/af/vfae031

Coghlan, S., & Parker, C. (2023). Harm to Nonhuman Animals from AI: a Systematic Account and Framework. Philosophy & Technology, 36(2). https://doi.org/10.1007/s13347-023-00627-6

Fjodorova, N., Novič, M., Venko, K., Rasulev, B., Türker Saçan, M., Tugcu, G., Sağ Erdem, S., Toropova, A. P., & Toropov, A. A. (2023). Cheminformatics and Machine Learning Approaches to Assess Aquatic Toxicity Profiles of Fullerene Derivatives. International Journal of Molecular Sciences, 24(18), 14160. https://doi.org/10.3390/ijms241814160

Hartung, T. (2023). Artificial intelligence as the new frontier in chemical risk assessment. Frontiers in Artificial Intelligence, 6. https://doi.org/10.3389/frai.2023.1269932

Igarashi, Y., Kojima, R., Matsumoto, S., Iwata, H., Okuno, Y., & Yamada, H. (2024). Developing a GNN-based AI model to predict mitochondrial toxicity using the bagging method. The Journal of Toxicological Sciences, 49(3), 117–126. https://doi.org/10.2131/jts.49.117

Jeong, J., & Choi, J. (2022). Artificial Intelligence-Based Toxicity Prediction of Environmental Chemicals: Future Directions for Chemical Management Applications. Environmental Science & Technology, 56(12), 7532–7543. https://doi.org/10.1021/acs.est.1c07413

Mellor, D., & Beausoleil, N. (2015). Extending the ‘Five Domains’ model for animal welfare assessment to incorporate positive welfare states. Animal Welfare, 24(3), 241–253. doi:10.7120/09627286.24.3.241

Neethirajan, S. (2023). Artificial Intelligence and Sensor Innovations: Enhancing Livestock Welfare with a Human-Centric Approach. Human-Centric Intelligent Systems, 4(1), 77–92. https://doi.org/10.1007/s44230-023-00050-2

Rosati, A. (2024). Guiding principles of AI: application in animal husbandry and other considerations. Animal Frontiers, 14(6), 3–10. https://doi.org/10.1093/af/vfae045

Rudroff, T. (2024). Artificial Intelligence as a Replacement for Animal Experiments in Neurology: Potential, Progress, and Challenges. Neurology International, 16(4), 805-820. https://doi.org/10.3390/neurolint16040060

Serrano, D. R., Luciano, F. C., Anaya, B. J., Ongoren, B., Kara, A., Molina, G., Ramirez, B. I., Sánchez-Guirales, S. A., Simon, J. A., Tomietto, G., Rapti, C., Ruiz, H. K., Rawat, S., Kumar, D., & Lalatsa, A. (2024). Artificial Intelligence (AI) Applications in Drug Discovery and Drug Delivery: Revolutionizing Personalized Medicine. Pharmaceutics, 16(10), 1328. https://doi.org/10.3390/pharmaceutics16101328

Singer, P., & Tse, Y. F. (2022). AI ethics: the case for including animals. AI and Ethics. https://doi.org/10.1007/s43681-022-00187-z

Like (0)