Building a Future of Trauma-Informed AI: Why This New Toolkit Matters
As AI continues to shape how people access information, support, and services, it is critical that we design these technologies in ways that are safe, ethical, and trauma-informed. At Aimee Says, we know that AI can be a lifeline for survivors of domestic abuse, providing validation, guidance, and a sense of control at a time when those things feel out of reach. But we also know that AI, if designed poorly, can replicate harm, erode trust, and fail the very people it aims to serve.
That’s why we were excited to see the Trauma-Informed Algorithmic Assessment Toolkit, a new resource designed to help organizations ensure their AI-driven services are built with the needs of trauma survivors in mind. This toolkit is a step forward in establishing best practices for AI that interacts with people who have experienced abuse, oppression, and harm.
At Aimee, we’ve been working on trauma-informed AI since generative AI started grabbing the attention of the public, and we recognize the value of this toolkit in helping the broader field move toward safer, more survivor-centered AI. Here’s why it matters.
This toolkit is built around five key trauma-informed principles, which are essential to ensuring that AI systems truly help survivors rather than harm them.
Empowerment & Choice
Survivors of abuse often have their autonomy stripped away by their abuser. AI should restore control, not take it away further.
The toolkit emphasizes that AI should:
- Give users meaningful choices in how they engage.
- Respect user agency over their data and interactions.
- Avoid forcing interactions that feel invasive or disempowering.
Aimee ap users can control what they share, how Aimee remembers their story, and when they engage—without pressure or judgment.
Collaboration
The toolkit highlights the importance of involving survivors, advocates, and professionals in the development of AI tools. AI systems shouldn’t be built for survivors without their input; they should be built with them.
Aimee follows this approach by:
- Collaborating with domestic violence advocates and survivors to shape how Aimee interacts.
- Incorporating real-world survivor experiences to make responses validating and nuanced. When survivors tell us what they need, we respond.
To collaborate with us, please reach out via the link at the bottom of this page.
Trust & Transparency
Many survivors have experienced broken trust—from people, systems, or institutions that failed them. AI should build trust, not erode it further.
The toolkit urges AI developers to:
- Be upfront about what AI can and cannot do.
- Make interactions predictable, so users aren’t caught off guard.
At Aimee, we ensure that:
- Survivors always know how Aimee works—there are no hidden agendas.
- Aimee is transparent about when a human professional may be needed instead of AI support.
Safety
For survivors, safety isn’t just about data protection—it’s about emotional safety, validation, and avoiding retraumatization.
The toolkit encourages AI developers to:
Ensure AI responses don’t mimic abusive tactics (e.g., controlling, gaslighting, or minimizing experiences).
Prevent AI from giving dangerous or invalidating advice.
Design AI with built-in safeguards to reduce harm.
Aimee’s protects user data through adherence to best practices, and is built to be a safe space—free from judgment, surveillance, or coercion.
Intersectionality
No two survivors experience trauma in the same way. Factors like race, disability, gender identity, socioeconomic status, and cultural background all shape how people experience abuse and seek support.
The toolkit stresses that AI should:
Recognize and adapt to diverse survivor experiences.
Ensure that responses aren’t biased against marginalized communities.
Acknowledge systemic barriers that impact survivors differently.
At Aimee, we’ve built adaptive AI responses that take these factors into account—ensuring that every survivor feels seen and understood.
If you’re a survivor, advocate, or AI ethics professional, we’d love to hear from you:
How do you think AI should be designed to better support survivors?
Reach out below—we’re always listening.
