How can we
make educated decisions about
AI use in our work as progressive activists?

( A GUIDE FOR AI USE IN PROGRESSIVE MOVEMENT WORK)

Which AI systems are we
using, & who is behind them?

QUESTION 02

Do the values to the entities behind these systems align
with my orgs values?

QUESTION 03

What is the impact of using these systems? (cost)

As movement workers, it is crucial that we understand how AI tools are used in order to navigate an AI-driven future. We must also be aware of who is behind these tools so that we can make informed decisions about when and if we should use them. Furthermore, we need to explore how we can turn these tools against the oppressors while staying true to our core values.

How can we make educated decisions on when to use AI and machine
learning in our work as progressive activists?

As we learned in the AI Education section, using AI tools comes with a cost– the impact to the environment, data security risks, and the bias algorithms many of these companies use when training their systems. So, how can we as nonprofit and movement workers keep up with the rapid growth of AI tools while also ensuring we stay true to our foundational values?

QUESTION 01

For example, the Department of Homeland Security (DHS) has expanded its use of AI tools, including those developed by companies like OpenAI. While your organization may not directly use the same tools or work with DHS, it is still participating in a shared ecosystem. Many AI companies provide services to both private users and government agencies, meaning that widespread adoption of their tools helps sustain and grow companies that also contract with entities whose actions may harm immigrant communities.

In this way, using these tools is not entirely separate from those outcomes – it can indirectly support the same companies and infrastructures that enable them. This raises a critical question: if a company’s partnerships or practices conflict with your organization’s values, should you continue to adopt and rely on its technology?

Ultimately, this is not just a question of efficiency or convenience, but of alignment. Supporting a technology provider also means, in part, supporting the broader network of relationships and impacts tied to that provider. Organizations must consider whether that alignment reflects how they want their members and communities to be treated.

Protecting your members’ or customers’ information is essential for any organization. Depending on the type of data you store, you likely should not input that same information into a chatbot or other AI tool.

While some machine learning tools can be useful – for example, in categorizing field notes, improving grammar, or organizing groups; it’s important to carefully evaluate any tool before using it.

We strongly recommend that you never enter the following types of information into generative AI tools:

  • Citizenship or immigration status

  • Social Security numbers

  • Legal information (including criminal history or active warrants)

  • Information about minors or children related to members/customers

  • Credit card or banking details

Am i considering…?

    • Do I fully understand how this tool works? Is my data secure when using it? While AI systems can be complex, it is essential to understand how the tools you use are developed and operated.

      How does the company train its AI models? Many systems improve over time by learning from user inputs, often by default. This means your data may be used in ways you are not fully aware of or have not explicitly consented to. In some cases, you may be asked to grant access to additional information - for example, through platforms like Google Workspace.

    • Do I understand the environmental, social, and broader impacts of using this tool? Are there more sustainable or ethical alternatives available?

      Evaluating the full impact of AI use, including energy consumption, data practices, and community effects, is an important part of responsible decision–making.

    • AI “hallucinations” occur when a system generates incorrect or misleading information and presents it as factual. This poses serious risks, including the spread of misinformation and erosion of trust within teams and communities.

      Hallucinations can occur across all AI tools, including AI detection systems designed to identify AI-generated content. Research from Stanford scholars has shown that some of these detectors are both unreliable and biased, particularly against non-native English speakers. In these cases, human-written content has been incorrectly classified as AI-generated.

      This highlights the importance of critically evaluating AI outputs rather than accepting them at face value.

    • Does this tool and the company behind it align with my organization’s values? A tool may appear highly effective, but using it can also mean supporting and potentially contributing data to companies whose practices may not align with your mission.

      Careful consideration should be given to whether adopting such tools is consistent with your organization’s principles.

QUESTION 04

What information are we putting into these systems?

AIM TO SUPPORT groups that are working to create equitable, unbiased
or sustainable AI

Ecosia

A Berlin-based search engine and generative AI chatbot that uses its advertising revenue to plant trees in an effort to offset its carbon footprint. Ecosia aims to protect user data by opting all users out of personalized ads by default, and it encrypts searches for an additional layer of privacy.

    • Although Ecosia works to offset its environmental impact, its chatbot feature still requires significant energy to operate

    • The chatbot may produce false or misleading information

    • Partners with large tech companies such as Google and OpenAI

Change Agent

An AI model designed for nonprofits, advocacy, and organizing, positioned as a “mission-based alternative to Big AI.” It aims to be environmentally sustainable, estimating that roughly 250 users consume the same amount of energy as one refrigerator. According to its website, it does not use user input for training and is designed to mitigate harmful bias while centering social sector values and perspectives.

    • Relatively new, with limited publicly available information

    • Currently only available as a demo

  • Higher Ground Labs – AI Resource Guide for Progressive, Political,
    & Democracy Use Cases

    This resource provides an introduction to artificial intelligence for beginners, along with curated materials such as workshops, training programs, policy templates, and practical guides to support responsible
    AI use in political and democratic contexts.

  • Techtonic Justice – Decision Guide for Considering AI Use

    Developed to support nonprofits and organizations, this guide outlines
    key considerations for adopting AI, including evaluation frameworks, ethical concerns, and the potential impact of AI tools on the communities they serve.

  • NetHope - nethope.org

    This platform offers access to a wide range of educational resources, including webinars, toolkits, and workshops, aimed at helping nonprofit organizations effectively and responsibly integrate AI into their operations.

  • United Nations Environment Program – AI has an environmental problem—here’s what the world can do about it

    This article explains the fundamentals of artificial intelligence, examines its environmental impact, and presents potential strategies for reducing
    its ecological footprint.

  • IBM – The Impact of AI

    This piece provides a balanced overview of artificial intelligence, discussing both its advantages and its challenges, including economic, ethical, and societal implications.

EcoDataCenter

A company that provides climate-friendly data centers in Sweden, powered entirely by renewable energy sources such as solar, wind, and hydropower. To reduce waste, EcoDataCenter integrates with local energy systems to recycle excess heat generated by servers and IT equipment, using it to heat buildings in Falun, Sweden. During the summer, excess heat is redirected to support cooling systems.

    • Despite using renewable energy, large-scale data centers still consume substantial amounts of energy

    • Construction can create environmental impacts such as noise and light pollution

    • Long-term land use concerns, including resource-intensive building materials (e.g., cement, metal, wiring) and land allocation that may affect local communities

AI RESOURCES

AI POLICIES AND PROPOSALS

HOW CAN I LEARN MORE ABOUT AI?

Viro AI

Marketed as a “clean alternative” to ChatGPT, OpenAI, and other popular AI chatbots. Viro AI aims to offset its environmental impact by funding renewable energy projects for each user prompt. It also states that it does not store user input for training. However, rather than being a standalone AI model, it provides access to systems like Gemini, Claude, and ChatGPT in one platform.

    • Potentially confusing user interface

    • Relies on major AI providers such as Gemini and ChatGPT

    • Focuses on carbon offsetting rather than building fully sustainable infrastructure

  • We developed a “Movement Workers Questionnaire” to gather insights into how individuals in the progressive movement perceive the role of artificial intelligence in their work. Based on the survey findings, we created a guide that directly addresses participants’ key questions and provides educational support tailored to the challenges and concerns they identified. Please follow the link to access the survey analysis from the “Movement Workers Questionnaire.

  • A comprehensive list of resources and references used to inform
    this guide.

CITATIONS

  • EU General Data Protection Regulation (GDPR) 

    The General Data Protection Regulation (GDPR) is a law in the European Union that protects the personal data of people living in the EU. It requires businesses and organizations to get clear permission from individuals before collecting or using their personal information. The law also gives people explicit  rights over their data. These include the right to access their information, correct mistakes, limit how it is used, or have it deleted. In addition, the GDPR includes the right to data portability, which allows individuals to copy and transfer their data to another service in a simple, usable format.

  • State/Local Initiatives on Sustainable Data Centers

    This article analyzes the impact of both state and local policy proposals designed to promote the development of environmentally sustainable data centers. It further explores a range of examples from states and municipalities that are actively implementing “green standard” initiatives, highlighting their approaches to improving energy efficiency, reducing environmental impact, and encouraging sustainable infrastructure practices within the data center industry.

Movement Workers Survey Analysis

List of Works Cited