In our previous blog, we explored why social workers need to approach Artificial Intelligence with care. This second piece focuses on the how; offering practical, values-driven strategies for using AI in a way that strengthens (rather than undermines) the heart of social work.
These tips are inspired by Fairspace’s own AI Policy, particularly our Ground Rules, which were created to help our team use AI without losing sight of nuance or lived experience. Think of this as a gentle, human-centred guide, not to make you “more productive,” but to help AI support your work, your judgment, and your care for others.
AI should lighten the load, not set new expectations.
This is the first rule in Fairspace’s policy. AI is optional. It’s a support tool, not a performance standard. No one should feel pressure to work faster or be judged for preferring human thinking over convenience. Use AI only if it genuinely helps you create more space for empathy, reflection, and connection.
Start and finish with human thinking.
Fairspace’s ground rule: “The first input and the last output should be from us, not the AI.” What this looks like in practice is:
- Start with your own draft, notes, thoughts, or messy brainstorm.
- Let AI support the middle of the task: summarising, clarifying, structuring.
- End with your own review, judgment, and edits.
This protects your creativity, your voice, and your critical lens.
Be a critical curator: review AI with three key questions.
Fairspace uses a simple three-question reflection for any AI-generated content:
- Do I take ownership of this?
- Is the AI-bias minimised?
- Is it nuanced enough for the context?
If the answer to any of these is “no,” revise, or do not use the content.
Prompt with values and context (safely).
Tell the AI who you are, who the audience is, and what principles matter, for example:
trauma-sensitive, intersectional, inclusive, non-victim-blaming, accessible language. Ask for diverse perspectives or for the model to flag its own limitations.
When it’s safe, you can share non-sensitive context such as public materials, project descriptions, or learning objectives; this helps contextualise the output. But never upload sensitive materials. Fairspace’s policy is firm: No personal data, no trauma content, no confidential stories, no partner information, ever.
Share the burden: collaborate with colleagues
Fairspace’s collaboration guidelines remind us that you don’t have to make ethical decisions alone. It’s okay to ask: “Does this wording feel safe?” Peer review is always stronger than algorithmic review. So if something feels off, pause and ask for help
AI is never a substitute for professional supervision, team learning, or lived experience.
Be transparent. It builds trust.
Fairspace encourages to openly acknowledge AI-assisted work, both internally and externally. If AI played a meaningful role in creating something, say so. Transparency builds trust with colleagues, partners and communities, and helps everyone stay honest about where human judgement was applied.
Especially when working with partners or external teams, decide together if, how and when AI will be used in the project. Agreeing on this early prevents misunderstandings and helps everyone stay aligned on ethics, privacy and values.
A final reflection
Social work is human work. AI can support us by helping with admin, translation, or early drafts; but it should never lead us, rush us, or replace the emotional, relational parts of what we do.
By using AI intentionally, gently, and critically, we protect our clients, our partners, the communities we work with, and the integrity of our practice. These ground rules are a way of holding onto the values that make social work powerful.
Want to learn more about ethical AI at your organisation? Get in touch via zeynep@fairspace.co
AI Assistance Acknowledgement: This blog was partly drafted with AI assistance for early drafting and structure. All content was reviewed and refined by a human author.
