SHADE Newsletter 8th January 2026

Welcome to the forty third edition of the SHADE newsletter! 

SHADE is a research hub with a mission to explore issues at the intersection of digital technologies/AI, health and the environment. It is guided by a fundamental question: How should the balance between AI/digital enabled health and planetary health be struck in different areas of the world, and what should be the guiding principles?

The SHADE newsletter comes out four times a year. It takes an in depth look at selected topics, as well as highlighting new resources, events and opportunities in the SHADE space.

In this first newsletter of 2026 we highlight the ongoing and projected AI breakthroughs in health. We then turn the spotlight to the debate about the environmental harms of AI, before rounding up some of the latest tools to measure and mitigate these harms and those of digital health more generally. We conclude with a packed selection of resources, events and opportunities. We hope you enjoy it!

Please tell us what you like, what you don’t like and what you think is missing at [email protected].   

AI breakthroughs in Health

• AI is driving ever more breakthroughs in health research and health care - revealing distinct types of multiple sclerosis, making way for MS treatment to be based on underling biology not symptoms, speeding up drug discovery and development and developing diagnostic tools such as the Galleri multi cancer early detection blood test currently being piloted in the UK.

• Looking forward, recent advances in neuroscience offer inspiration for developing adaptive artificial intelligence. This could create adaptive agentic systems which would be more effective than existing AI agents in dynamic and uncertain environments.

• On AI agents more generally, 2026 may be the year that the first consequential scientific advances are made by AI, and AI is in the running to take over scientific research from humans by 2050.

The environmental harms of AI

• Alongside the positive news about AI in health and scientific research, questions over AI’s environmental sustainability are everywhere. At the local level, opposition to data centres is growing. This article in Tech Policy Press suggests this could put the brakes on the AI revolution in the US, and this report from rest of world highlights the growth of local pushback worldwide. Organisations and online hubs, such as DataCentreBoom! in Latin America, are being set up to support communities and local authorities concerned about data centre deployment in their area.

• Better environmental disclosure from tech companies is becoming increasingly urgent says Alex de Vries-Gao in this paper in Patterns. This includes the need to distinguish AI from non AI workloads so the rapidly increasing environmental impact of AI systems can be responsibly managed.

• Meanwhile, from Data and Society, Turning the Tide: Climate Action In and Against Tech by Tamara Kneese uncovers how climate conscious tech workers are adopting an organiser - rather than a lab - mindset to advocate for social change rather than ‘technical tweaks’. Alongside this, the report also acknowledges the ‘dire need for producing more holistic evaluation frameworks that consider quantitative energy and water use metrics’.

Tools

• Firstly, Beyond Counting Carbon: AI Environmental Assessments Struggle to Inform Net Impact Decisions. This pre-print from the University of Bristol highlights just how sensitive AI environmental impact methodologies are to the principles and parameters that underpin them, ‘highlighting the need for more transparent, consistent, and AI-specific approaches’.

• The Green Software Foundation (GSF) have been asking how we reliably measure the carbon footprint of AI and, after an extensive collaborative process, they have come up with an answer in the form of a Software Carbon Intensity (SCI) for AI specification. This captures AI’s complete environmental impact across its entire lifecycle. The GSF intend to provide more supporting documentation for the methodology in 2026.

• Picking up on the AI takeover of scientific research, this article from Nature Communications proposes a triadic framework combining human regulation, agent alignment, and environmental feedback to provide ‘a more comprehensive foundation for safe AI scientist development’. The authors take into account risks AI scientists pose to nature, human health and social and economic stability and recognize that ‘no single technical solution can address all potential vulnerabilities’. This framework is one of nine just added to MIT’s AI Risk Repository.

• Finally, Quantifying the global eco-footprint of wearable healthcare electronics: This paper describes the full-spectrum environmental impact metrics generated for glucose, cardiac and blood pressure monitors and diagnostic imagers. This review article from Nature discusses all the issues addressed in the study which demonstrates that a more sustainable approach to the electronics components offers more promise than focussing on recyclable or biodegradable plastics.

Resources, Events and Opportunities

• Join us for an in person event in Oxford on Tuesday January 13th at 2pm. SHADE co-director, Federica Lucivero is chairing a conversation on Tools for environmentally sustainable health research: opportunities, challenges and open questions. Registration is required.

• The CHI conference on Human Factors in Computing Systems is the leading international conference on Human-Computer Interaction. Taking place in Barcelona from April 13th to 17th, it includes a workshop the goal of which is to lay the foundations for ‘building a community of HCI researchers interested in mitigating the environmental impact of AI and accordingly bring methods from our inherently interdisciplinary domain that go beyond solutionist narratives’. Find out more and read the call for participation. The deadline for paper submission is February 14th.

• Looking to change your diet in 2026? This podcast from the One Health Trust looks at a new report from the EAT-Lancet Commission which lays out how everyone can eat better to improve the health of the planet and themselves. It details the benefits of the Planetary Health Diet.

• Do you want to develop or work with early warning systems for climate-sensitive diseases? Register for episodes of this Advanced Webinar Series on Spatiotemporal Modeling of Climate-Sensitive Diseases from DHIS2. While open to everyone, these webinars introduce advanced technical concepts and assume prerequisite knowledge. You’ll get the most value if you have some experience with Python or R programming and some background in epidemiology, statistics, or data science.

• HISP network to support 7 African countries with DHIS2 Climate & Health tools. Through the Global Fund Climate x Health Catalytic Fund, HISP UiO and local HISP groups will help countries strengthen national information systems for climate-informed health programming and disease surveillance.

• Global perspectives on infectious diseases at risk of escalation and their drivers from Nature Scientific Reports warns of ‘creeping catastrophe’. In one of the largest global studies of its kind, this Wellcome commissioned research identifies three drivers - climate change, inequality and AMR - behind an escalating health crisis. The research involved thousands of health professionals across 151 countries. Oxford reports on reaction to the results of ‘this phenoimenal study’.

• The Indian Ocean disaster is a climate tragedy — and needs more attention says Nature, highlighting that three tropical cyclones in late November were almost as destructive as the 2004 Indian Ocean tsunami. The article calls for more attention ‘not only to apply pressure to speed up rescue and recovery, but also to recognize the planetary-scale crisis that humanity is currently facing’.

• The complexity of air pollution toxicity is illustrated in this study of short term exposure to ultra fine particles in London and the West Midlands. The analysis concluded that there was no evidence that such exposure is associated with nonaccidental, cardiovascular, or respiratory death, ‘emphasising the need to refine our understanding of which particulate metrics are most harmful’.

And finally, Climate AI’s dataset of the month is the Newfoundland Marine Refuge Fish Classification Dataset (N-MARINE). As ocean warming accelerates, this annotated data facilitates the monitoring of marine ecosystems using machine learning - an approach considered to be less invasive and more accurate than trawling .

We hope you have enjoyed this newsletter. If it has been forwarded to you, and you would like to receive future editions, you can subscribe here