ACORN Lab AI Policy 

As members of the ACORN Lab community, we are committed to supporting each other in our education, learning, teamwork, and career success. With regard to the use of Artificial Intelligence (AI), we commit to:

• Writing without the use of AI. We value building the challenging skills of scientific writing and communication so that we can be effective scientists and leaders in the field. We recognize that no writing is perfect and that we will continue to build our writing skills all throughout our careers, so we commit to giving and receiving constructive feedback from each other as we learn to write more clearly and impactfully. All peer-reviewed publications, conference abstracts, and grant materials from the ACORN Lab will be written by us.

 

• Being mindful of what information we put into AI tools. We never share any data from our research or any information about research participants or students with AI in compliance with the strict IRB, HIPAA, and FERPA security and privacy policies that we adhere to. 

 

• Supporting artists to help us create artistic materials. Our lab’s logo and other branding materials have been thoughtfully and collaboratively developed by artists from the ACORN Lab family. We commit to continuing to support artists rather than use AI to generate artistic material. We also commit to developing our creative skills in generating our own figures and schematics for our peer-reviewed publications and other research outputs. 

 

• Minimizing our use of AI. As scientists, we are conscious of the impacts of AI usage on the Earth’s climate and its disproportionate impacts on marginalized communities (See below: “Resources on Environmental Impacts of AI Use”). We aim to maximize our opportunities for scientific learning and education while minimizing harm. To this end, we commit to supporting each other in our learning of new materials and skills. We commit to trying things ourselves, continually seeking help from our teammates, and leveraging the many tools and resources at our disposal to minimize AI usage (See below: “Tips & Tricks”). 

 

• Recognizing our different levels of comfort with AI usage. We recognize that different team members may feel different levels of comfort with the use of AI for different purposes. We commit to respecting one another’s comfort levels and will never impose requirements that a team member must use AI. We commit to sharing and discussing our AI usage openly so that we can learn from one another and support each other. 

 

• Documenting our use of AI transparently. We commit to documenting any AI usage transparently in laboratory notebooks and other forms of documentation.

 

• Developing and applying critical thinking skills in evaluating AI outputs. We recognize that AI tools often provide outputs that are incorrect, biased, or misleading. As scientists, we commit to applying our caution and critical thinking when interpreting and evaluating AI outputs rather than treating output as fact. We aim to maximize our use of primary sources for information and to carefully evaluate the validity of AI-generated information. 

 

• Learning at a balanced pace. We believe that it takes time to learn the most important skills for scientific research, including scientific writing, computer programming, technical methods, and idea brainstorming. We recognize that a major advantage of using AI tools in scientific research is to accelerate the generation of research outputs, but that it takes slow and steady practice to learn things well. We as a team are committed to moving at a balanced pace, even when that means going slower to allow time for the messy process of human learning to take place. We commit to supporting each other in our learning and career growth, celebrating our mistakes as stepping stones on the path to success, and building self-confidence in our ability to learn new/challenging things. 

 

• Continuing our discussions and updating our guidelines. As the technological landscape continues to evolve and new community members join the team, we commit to continually learning, discussing, and updating our guidelines.

 

~~~~~~~~~~~~~~~~~~ Resources on Environmental Impact of AI Use ~~~~~~~~~~~~~~~~~~

Environmental Impact of AI Use:

• Data centers are physical facilities that store computing machines or servers, and related equipment. They are said to be the “backbone of the digital economy” and are expanding rapidly to meet AI demand. 
  • There are approximately 12,000 data centers globally, with 5,400 living in the United States. Data centers dedicated to powering AI are increasing at an exponential rate.
  • New facilities are being built daily and at a rapid pace, often near heavily populated, disadvantaged, or minority communities. 
• One singular large data center can consume as much electricity as 100,000 homes. “Hyperscale” data centers that are now being built are expected to consume 20 times that amount.
  • These centers are online and in use all day, every day, without interruption. An enormous amount of energy is needed to keep the computers on and clean water is needed to cool these systems down so they don’t overheat.
  • This much energy consumption releases significant amounts of carbon dioxide (CO2) and other greenhouse gases into the atmosphere, which contributes to climate change.
• Large data centers consume up to 5 million gallons of drinking water per day for cooling purposes. The global AI demand in 2027 is projected to account for 4.2–6.6 trillion liters of water. That’s enough water to fill 1.5 billion Olympic size swimming pools.

 

Impact of Data Centers on Local Communities:

• The large energy consumption contributes to increasing energy demand and rising electric bills paid for by the average American.
• Data centers require clean water for evaporative cooling, which is straining local water supplies. 
• Data centers create continuous air pollution and greenhouse gas emissions, increasing rates of serious health issues, such as asthma or heart disease.
  • A prime example of this is in Memphis, TN, where residents have reported a constant gas smell and worsened air quality.
• Noise pollution from generators and construction have been reported to cause disruptions in sleep, headaches, and lower quality of life. 

 

Environmental Policy:

• The Environmental Protection Agency (EPA) is primarily responsible for setting and enforcing environmental limits and regulations around air pollution, water quality, hazardous waste management, energy production, etc. 
  • This agency is being defunded, downsized, and facing major policy shifts under this current administration.
  • The “endangerment finding” under the Clean Air Act, which concluded that greenhouse gas emissions endanger public health, was officially repealed in February 2026.
  • This rule allowed the EPA to regulate facilities, processes, and pollutants that contribute to increases in greenhouse gas emissions.
• The significant rollback of regulations is detrimental, especially in combination with the massive rise in hyperscale data centers that are using enormous finite resources.

 

~~~~~~~~~~~~~~~~~~~~~~ Tips & Tricks for Minimizing Harm ~~~~~~~~~~~~~~~~~~~~~~

• Leverage available resources to answer your questions: talk to colleagues/mentors and utilize other online resources (e.g. PubMed, NeuroStars, Stack Overflow, Github).
• Reduce processing footprint by writing comprehensive queries (e.g. asking several questions within one prompt to reduce back and forth queries).
• Test code in small local batches before requesting large computational resources.
• Test simpler computational models before running more complex models (also a useful scientific best practice!).
• Favor more sustainable AI platforms (e.g., Ecosia AI, Viro AI, Open WebUI, Ollama); reduce the use of generative AI tools which require more power.
• Consider scale, usage, intention, value, and available alternatives before leveraging large computational resources.
• Practice reading scientific research articles that are outside your comfort zone without AI summaries; share papers with each other/discuss in journal club to facilitate learning together and learning from each others’ reading practices.