MakeWay does not provide or officially sanction AI tools for staff use. However, we recognize that many staff are using them independently, and we want to help you do so responsibly. We recognize the importance of having a clear organizational position on AI and are working toward a more comprehensive policy. Until then, these guidelines are intended to support consistent and responsible use and do not replace existing MakeWay policies. Like any technology, appropriate usage must fall within all of MakeWay’s policies. These guidelines apply to a range of AI tools, including but not limited to tools such as ChatGPT, Microsoft Copilot, Google Gemini, Claude, etc...

Used well, AI can be a useful tool for brainstorming, summarizing information, improving drafts, and supporting routine tasks. However, it cannot understand relationships, organizational culture, community context, or the lived experiences of the people we work with. Those require human judgment, reflection, and meaningful engagement with others.

AI does not replace human interactions, judgment, expertise, or relationships. It does not replace the guidance, support, and direction available from your peers, managers, or subject matter experts. When in doubt, talk to real qualified people.

Please read these guidelines in full before using AI tools for your work at MakeWay. Please also note, what you do with AI tools is subject to ALL MakeWay Policies.

Sharing Private Information

Do not request or share personally identifiable information, financial details, or any other Sensitive (Confidential or Restricted) information with AI. Protect user privacy and adhere to MakeWay’s policies and legal requirements.

Funder and partner restrictions

Some grant funders or partners explicitly prohibit the use of AI in applications, proposals, or other deliverables. In those cases, all materials must be prepared entirely without AI assistance. When in doubt, confirm whether AI use is permitted before proceeding.

Sensitive personal situations

Do not use AI for tasks involving the evaluation, assessment, or management of individual employees. MakeWay’s People Engagement and Learning (PEL) team does not use AI for recruitment, performance management, employee engagement, employee relations, or any employment decisions. Personal and confidential employee information must not be entered into AI prompts. PEL members will receive appropriate training before being asked to use any AI tools in their work.

Inappropriate Content

Do not use AI to generate or promote content that is offensive, hateful, discriminatory, or violent, or that otherwise conflicts with MakeWay’s code of conduct, policies, or applicable laws.

Emergencies

AI tools are not designed for emergency or life-threatening situations. Contact appropriate emergency services or qualified professionals directly.

Unqualified Advice

Do not use AI to provide legal, medical, financial, or other professional advice beyond your expertise. Refrain from offering definitive recommendations or making critical decisions solely based on AI’s responses.

AI Notetakers

Do not invite third-party AI Notetaker bots (e.g. Otter, Fireflies, Fathom, Read) to MakeWay meetings. Inviting AI Notetaker bots is not permitted.

AI Co-Working Models

Do not grant access to Folders on your device. Granting access to broad folder sections can create vulnerabilities and security risks in tools that you may not be aware of. Sharing single files within the boundaries shared in these guidelines and policy is permitted, but giving access to Folders and broad sections of your device can compromise the security of MakeWay’s operations.

Indigenous data - including data created by, collected about, or held on behalf of an Indigenous nation, community, or governing body - such as community, cultural, or traditional knowledge is subject to Indigenous Data Sovereignty principles (OCAP - Ownership, Control, Access, Possession). This applies whether the data originates directly from the community (e.g. traditional knowledge, community-generated datasets) or is data MakeWay has collected about a community or partner (e.g. grant reporting, program data). In both cases, the community retains authority over how its data is used, including in AI tools.

By default, do not enter Indigenous data into AI tools in any form, anonymized or not. If you have authority from, or documented consent from, the relevant community's governing body or other appropriate authority as determined by the Indigenous partner(s) you are working with, to use their data with AI under OCAP principles, consult your manager before proceeding so the use can be considered and documented. MakeWay does not yet have finalized guidance in this area.

AI tools are trained on large datasets that reflect human-generated content, including the biases present in that content. This means AI outputs can reflect or reinforce racial, gender, cultural, and other biases, sometimes in subtle ways. Before sharing AI-generated content, review it with this in mind. If you notice a response that seems biased, stereotyped, or not in line with MakeWay’s values or culture, correct it before passing it along. 

Disclosure

Do not present AI-generated content as purely human-generated in contexts where transparency matters, such as official reports, research studies, public communications, or for tasks or activities that require high human oversight and relational context. Use judgement, and where appropriate add a brief note indicating AI was used in drafting (e.g., "This document was drafted with the assistance of AI and reviewed by MakeWay staff"). If you're unsure, ask your manager. Internal working documents and brainstorms do not require this disclosure note but may be appreciated depending on the context, content, and people. If a funder, partner, or other party explicitly asks whether AI was used, always disclose this honestly.

AI-generated text and images may also raise copyright or intellectual property questions. Treat AI outputs as draft material that always requires review, and where appropriate, proper sourcing or adaptation.

AI tool settings

Regardless of which AI tool you use, confirm that the setting to allow your data to be used for AI training is turned off. Reputable tools offer this in ‘account’ or ‘privacy’ settings. Turning this setting off does not eliminate all risk, data is still transmitted and processed. This is why sensitive data should always be anonymized even when this setting is enabled.

Before entering any information into an AI tool, consider how sensitive it is. Always follow MakeWay policies and never enter Sensitive (Confidential or Restricted) information into an AI prompt or tool.

Low sensitivity – generally fine to include:

• Publicly listed information
• General process or workflow descriptions
• Aggregated or statistical data with no identifiable individuals
• Draft text that does not reference real people or confidential details

Medium sensitivity – use with care, anonymize or adjust where necessary:

• Internal organizational information not intended for public audiences
• Grant application content or strategic planning documents
• Descriptions of internal processes, decisions, or challenges that could be sensitive if shared externally
• Internal communications such as meeting notes, team updates, or staff announcements

High sensitivity – never share OR always anonymize, regardless of tool or settings:

• Personal contact information (home addresses, personal phone numbers, personal email addresses)
• Financial records, donor giving data, or banking information
• Health or medical information
• Employee information (performance, compensation, sensitive interactions)
• Vendor or partner information, including contract details or ongoing negotiations
• Descriptions and details of interpersonal situations or team dynamic
• Any information subject to a legal confidentiality obligation

Anonymizing means replacing identifying details with placeholders before the content enters the AI tool. For example, replace a home address with [address], or use [name] instead of a donor name. Re-link real data only after you have exported the AI output and are working outside the tool.

Writing correspondence with AI

Being a primary point of contact for most staff, email correspondence is highly relational. Tone and context matter. We don’t recommend letting AI write your emails. However, using it as a tool to draft, organize thoughts, or frame content can sometimes be helpful. Always use your own voice, consider who your audience or recipient is, and consider their comfort in the use of AI before deciding to use it for correspondence.

If you do use AI to help draft content, never share sensitive details about anyone: their contact information, private details, health situation, compensation, or anything you would not want to appear outside the organization.

AI tools are often fluent and confident sounding, even when they are wrong. Before sharing or acting on AI-generated information, always verify the output.

Lower risk:

• Brainstorming, generating ideas, or exploring options
• Drafting and formatting text you will review and edit
• Summarizing content you already understand

Higher risk:

• Specific facts, statistics, dates, or citations
• Information about specific people, organizations, or events
• Anything that will be shared externally or used in official documents

Cross-reference with reliable sources and apply the same critical thinking you would to any other information. If a topic is outside your expertise, escalate to someone qualified rather than relying on AI output.

AI generates responses based on patterns in data rather than an understanding of your specific context. While it can produce useful suggestions, it cannot exercise judgment or understand the nuances of our work.

You are responsible for everything you produce, whether or not AI helped you create it. Review and verify AI outputs before acting on them or sharing them. Ensure they align with MakeWay’s mission, values, and standards.

If a task becomes complex, sensitive, or requires expertise beyond your own, escalate it to a qualified person. For language translation of official public content and information, consult a human translator. AI is not a substitute for professional advice in legal, financial, medical, or other specialized areas.

Day to day tasks don’t need powerful AI models. Larger, more capable models use more energy and water, so be conscious of the model you’re using for the task at hand. Sometimes using more advanced models for simple tasks actually slows down results due to over processing and doesn’t provide any additional benefit.

If possible, set your tool’s default to a low or medium processing model, and only elevate the model for higher processing when needed.

Be cautious about which AI tools you use. Avoid tools from unverified or unknown sources that may pose security risks to MakeWay’s network or data. If you’re unsure whether a tool is appropriate to use, contact IS at [email protected] before proceeding.

AI tools and best practices are evolving quickly. If you choose to use AI tools as part of how you work, we encourage you to stay informed about changes that may affect privacy, accuracy, or appropriate use. Always follow MakeWay policies, use these guidelines, and reach out to IS if you have questions.

Questions? Contact the Information Systems team at [email protected].

These Guidelines were drafted by MakeWay staff members with editing assistance of an AI tool. They were reviewed and finalized by the MakeWay AI Pod members.