AnimalHarmBench debut, digital sentience funding, AI-animal contractor opportunities
All the latest on AI and nonhumans in June 2025
Welcome to the tenth edition of AI for Animals, the newsletter bringing you the latest on AI’s potential impacts on non-humans.
We’re trialling a more reader-friendly layout this month! Help us make the newsletter better by sharing your feedback or submitting a content request for future editions. You can subscribe to this newsletter here.
Thanks for reading!
Max, Constance, Chloe, and the AI for Animals team
Index
🌟 Highlights
A quick roundup of key updates, must-reads, and standout moments from this edition of the newsletter.
From 📢 Community News: Arturs Kanepajs presented the first ever standardized, empirical benchmark measuring AI harm risk to animals at the 2025 FAccT conference (the Association for Computing Machinery Conference on Fairness, Accountability, and Transparency). You can read more about the AnimalHarmBench benchmark in Arturs’ X post and in this Faunalytics summary.
From 💼 Opportunities: Longview Philanthropy, Macroscopic Ventures, and The Navigation Fund have opened three funding streams for work on digital sentience—covering applied projects, research fellowships, and career transitions. Apply by July 9th.
From 💼 Opportunities: Rethink Priorities and AI for Animals are seeking new contractors and volunteers! Rethink Priorities is looking for support for AI strategy and governance projects at the intersection with animal welfare, global health, and digital sentience. AI for Animals is searching for an executive assistant, Slack community volunteers, and expressions of interest for other higher level roles.
📢 Community News
🇬🇧 AI, Animals, and Digital Minds London 2025 was a success! Check out our highlights video. We are working hard to get the session recordings and full retrospective ready soon.
✨ AI for Animals is planning a rebrand to Sentient Futures, reflecting our focus on digital beings as well as biological animals. We will be hosting a conference and unconference in New York sometime from October 3-10 – stay tuned for more info shortly.
We are welcoming 3 new members to the team, Jay Luong and Damin Curtis who will be focusing on course and fellowship development and Sam Chapman who will be focusing on our Slack community. More to come about their roles later.
🐔 Kevin Xia has written several EA Forum posts about AI and animals:
Transformative AI and Animals: Animal Advocacy Under A Post-Work Society: In a post-work society shaped by AI and universal basic income, animal advocacy may gain volunteer capacity but struggle against automated opposition, cultural inertia, and moral complacency.
Asymmetries, AI and Animal Advocacy: Though AI can empower all sides, advocates can leverage asymmetries – efficiency, cooperation, moral framing – to outpace industrial agriculture.
Anchoring AI and Animals: Advocates should focus on influencing AI actors, growing support for alternatives, and exploiting strategic asymmetries over industry.
📹 Adrià Moret ran a webinar on "Including Animal and AI Welfare in AI Alignment" as part of Rethink Priorities’ Strategic Animal Webinars series.
🔍 Elizabeth Álvarez writes that organizations should adopt Answer Engine Optimization to ensure their ideas are cited accurately in AI-generated responses, improving reach, credibility, and impact as more people rely on chatbots instead of search engines.
🌐 Constance Li offers guidance for aspiring ‘super connectors’, emphasizing clear purpose, deep understanding of people, selective introductions, and stewardship of ‘whisper networks’ to build trust and maximize impact through thoughtful, high-value connections.
🔮 Magnus Vinding argues that even if we can’t predict the long-term effects of our actions on suffering (aka we are “clueless,” it still makes sense to focus on reducing suffering in ways we can realistically assess and influence, rather than giving up on trying altogether.
🧠 Interactions with Nirvanic Consciousness Technologies, a company aiming to create conscious AI using quantum computing, starts a dialogue about its approach to potential AI suffering.
🖥️ Chris Kroenke proposes a framework to systematically evaluate and improve large language models’ ethical biases toward animals by measuring how models value different species, comparing outputs to expert and public views, creating interventions like fine-tuning and constitutional AI, and continuously tracking progress to promote more compassionate AI systems.
✍️ Max Taylor shared a draft strategy for integrating AI into research, based on his work at Animal Charity Evaluators.
💼 Opportunities
Animals in the Room and Radboud University (Netherlands) are recruiting a PhD candidate to design and test new models that give non‑human animals a voice in decision‑making. Apply by July 6th.
Manifund regrantor Marcus Abramovitch is actively looking to fund ideas that combine AI safety with animal welfare. Pitch him directly on the AI for Animals Slack; funding is awarded on a rolling basis.
The Veit Lab for Animal & AI Sentience at the University of Reading welcomes post‑docs, PhD, and Master’s students keen to explore consciousness, welfare, and policy for both animals and AIs. Expressions of interest are accepted year‑round; apply here.
📅 Upcoming Events
Technology in Wildlife Welfare Workshop, July 4, Nottingham
Can AI Replace Bryant Research?, July 10, Online
Compassionate Future Summit, August 21–25, Berlin
European Federation of Animal Science Annual Meeting (focussing on the theme of “Future-proof livestock farming”), August, 25–29, Innsbruck
AI(Live), 23 September, London
EA Global: New York City 2025, October 10–12, New York City
📚 External News
🐔 Farmed chickens
WATTPoultry writes that U.S. poultry producers are rapidly adopting AI and robotics to boost processing efficiency, reduce labor costs, and meet rising demand while improving data collection and energy use.
At the Animal Agriculture Alliance Stakeholders Summit, a former Google executive outlined how generative AI tools like ChatGPT are being embraced by poultry marketers to accelerate campaign development, simplify data, and tailor content, with the key message being that those who don’t adopt these tools risk falling behind competitors who do.
At a poultry farming panel hosted by Poultry Business, industry experts argued that AI tools can help farmers detect problems sooner, improve bird welfare, and run operations more efficiently, but warned that high costs, data issues, and trust concerns mean producers should start with small projects, build good data, and learn how to use AI effectively.
A new report urges the UK poultry sector to invest quickly in training workers to use automation and AI, warning that without more skills in data analysis and digital tools, the industry risks falling behind as technology reshapes farm jobs and flock management.
🐟 Other farmed animals
Norway hosted the first AI-focused aquaculture conference, Aqkva AI, highlighting how artificial intelligence is rapidly advancing fish farming by enabling data-driven decisions across the value chain, with major producers like Lerøy and Mowi now leading digitalisation to boost efficiency and biological performance.
Ace Aquatec and Aquascot are working together to use AI cameras in Scottish salmon and trout processing plants to measure fish size and quality in real time, helping reduce manual checks, improve tracking, and ensure consistent product standards.
MACSO, a specialist in AI disease-detection sensors, and Unigen, a maker of on-site AI computing hardware, have partnered to deploy AI-powered audio sensors and on-premise inference servers in pig farms, enabling earlier detection of respiratory illnesses—responsible for 60% of global pig deaths—and reducing antibiotic use by identifying disease days before humans can.
🍔 Alternative proteins
The Bezos Earth Fund awarded nine of 24 initial AI Grand Challenge grants to projects using AI to advance sustainable protein production, each receiving $50,000 to develop alternatives to conventional animal agriculture and improve food system efficiency.
In 2024 and early 2025, AI attracted unprecedented capital (over $100 billion), diverting funds from alternative proteins and other sectors, which saw their weakest deal activity in a decade. FoodDive provides further context about this trend.
Basecamp Research has launched BaseData, a dataset containing 9.8 billion new protein sequences from over a million newly discovered species, to vastly expand biological training data for AI models, aiming to improve protein design by offering 10 times more sequence diversity than all public databases combined. However, some experts warn that without more context about these organisms, such vast genetic datasets may have limited practical value.
🦘Wild animals
Shannon Ray from Wild Animal Initiative writes that AI-driven unsupervised translation models could soon decode wild animal vocalizations, enabling two-way communication that may transform welfare science but also raises ethical risks of misuse, privacy breaches, and unintended harms.
Researchers are using AI to rapidly analyze massive wildlife datasets and improve conservation decisions, but critics warn this shift risks sidelining field skills, reinforcing Western biases, and diminishing Indigenous ecological knowledge.
Oregon State University researchers showed that training AI models on fewer, carefully selected trail camera images focused on one species and diverse local environments improved wildlife identification accuracy to nearly 90% while reducing data, computing needs, and energy use.
An article in Time describes how AI is being used to monitor and protect oceans by powering tools like digital twins that model fishing impacts, smart cameras that optimize aquaculture feeding, and video systems that detect illegal fishing, aiming to curb overfishing and pollution.
🗣️ Understanding animals
A Guardian article argues that AI is rapidly advancing efforts to decode whale and dolphin communication, promising interspecies dialogue but raising profound questions about whether humans can truly grasp or ethically transform animal experience without imposing our own perspectives.
The Earth Species Project community is holding dialogues on ethical safeguards for interspecies communication technology, inviting people to attend meetings and provide written contributions on this topic. Join the #interspecies-communication channel on the AI for Animals Slack to stay up to date with these discussions.
🐶 Companion animals
China’s AI pet industry is rapidly expanding, offering emotionally responsive robots like ROPET and BabyAlpha as alternatives to live animals, with applications ranging from companionship to mobility assistance for the visually impaired and strong domestic and international sales growth.
A piece in Kinship argues that AI tools analyzing animal facial expressions could improve welfare by detecting pain or distress more accurately than humans but may risk weakening owners’ attentiveness and emotional connection by outsourcing understanding of pets’ signals to technology.
A BVA survey found 21% of UK vets now use AI daily, mainly for radiography and lab diagnostics, citing faster data interpretation as a key benefit, while most expressed concerns about over-reliance, loss of human skills, and lack of contextual judgment.
Fi has launched a new AI-powered dog collar that reportedly tracks movements, health habits, and behaviors like barking or scratching with 80% accuracy.
🐭 Animal testing
Vivodyne raised $40 million to scale its AI and robotics platform for testing drugs on lab-grown human tissues, aiming to replace animal models and launching a fully robotic lab in San Francisco to expand high-throughput, clinically predictive research.
🌲 Evergreen Resources
If you’re looking to get up to speed with the basics of AI’s impacts for animals, we recommend:
Joining the AI for Animals Slack
AI ethics: the case for including animals by Peter Singer & Yip Fai Tse
Harm to Nonhuman Animals from AI: A Systematic Account and Framework by Simon Coghlan & Christine Parker
What AI Could Mean for Animals by Max Taylor
That’s it for this edition - feel free to get in touch by sharing your feedback or submitting a content request for future editions!