Disaster Recovery Is a Full-Time Job Nobody Trains You For

67% Are Still Displaced. Here's What Might Actually Help.

Two weeks ago, a glacier tip broke off in Nepal after the region experienced some of the warmest temperatures on record, triggering a flood that’s killed over a thousand people, with more than five thousand still missing, injecting stress and trauma into the entire population, with no end in sight. Meanwhile, it feels like the rest of us have already moved on because it’s disappearing from the headlines. We know this pattern: a climate disaster hits, the world pays attention briefly which stimulates aid and compassion, supports that then dry up as the catastrophic event loses salience when the media disappears, leaving communities to trudge through tidal waves of hardship largely on their own as they attempt to rebuild their lives, until… WHAM! Another disaster hits elsewhere, and the cycle repeats. The original survivors still need lots of help while new survivors enter the fold.

At Stanford, we’ve been trying to interrupt the cycle by ideating and testing more effective disaster recovery tools. I’ve been working with Dr. Adrienne Heinz, a clinical psychologist and disaster mental health specialist, pairing disaster-recovery science with AI to support survivors after the press leave and aid disappears. My team and I are collaborating with her nonprofit, The After Collective, and their Disaster Recovery Coach, a text-message-based AI tool that gives survivors the information, skills, and trusted guidance they need to navigate recovery. This week I’m sharing the findings from a new pilot study they conducted with 115 survivors from the 2025 Eaton and Palisades wildfires.

I have real reservations about AI right now, including its ownership models and the risk for what I think of as industry “bot-washing” of the tech’s detrimental societal effects by funneling philanthropic dollars towards university researchers so they can innovate with AI, which supports an appearance of ethical neutrality. If the trend grows, it’ll echo of what the fossil fuel industry has done with how it funds research at R1 universities for years. But those concerns are parked for this post, and I’ll have to return to the trouble of them in future writing. The question I’m exploring here is: can AI keep survivors from falling into disillusionment once the world moves on? Early signals say yes, quite possibly.

The Weight

The After Collective’s pilot study found that more than a year later, 67% of the Eaton and Palisades wildfire survivors in this study were still displaced, and 73% identified emotional or mental health stress as their biggest recovery challenge. 61% were still wrestling with insurance claims and rebuilding paperwork; 52% were in real financial strain; a quarter wanted mental health support they couldn’t get; and 30% rated their emotional wellbeing as poor or very poor. Again and again, people described recovery as a second full-time job: insurance claims, temporary housing, financial decisions, rebuilding, paperwork, and deadlines. All of it layered on top of grief, trauma, parenting, work, and trying to rebuild some sense of normal life.

There is a cruel mismatch here. Trauma disrupts attention, memory, planning, and decision-making, yet recovery demands enormous amounts of exactly those cognitive capacities. We ask people to become expert case managers for their own catastrophe at the moment their brains and bodies are least equipped to do it.

We are entering a particularly consequential moment. Climate disasters are placing more people into these long recovery journeys, while the systems responsible for helping them are strained and the future of federal disaster support (for example, in the US, with FEMA) is increasingly uncertain. At the same time, AI is advancing extraordinarily quickly. The recovery systems that use it must be shaped by people who understand what disaster survivors actually experience and how communities find their footing again, over the long term.

The Work

Survivors didn’t need another enormous library of disaster resources. They were already drowning in information. They needed help answering a firehose of questions: What applies to me? What do I do first? Can someone help me get this thing done? They needed coping skills to endure the emotional toll and support their families. My kids have so much separation anxiety. I keep having nightmares and can’t sleep. I’m drinking more now because of the stress.

Recovery does not divide neatly into logistical and emotional challenges, so recovery support shouldn’t either.

The goal of the coach is not to replace therapists, case managers, or community organizations. It is to return agency, reduce overwhelm, and take some of the burden off people so they have more capacity for the human parts of recovery.

Know what to do → find what you need → get difficult things done → protect your psychological capacity to keep going.

AI may be especially useful for tasks that require precious recovery energy: searching through dozens of websites, organizing time-sensitive information, figuring out which resources apply to them, filling out forms, breaking a complicated task into manageable steps, or keeping track of what comes next. It can also help someone cope with stress at 2am on a sleepless night, using evidence-informed protocols for disaster like Psychological First Aid and Skills for Psychological Recovery.

69% percent of the survivors surveyed said they’d want AI-powered support like this; general availability of the free Disaster Recovery Coach is planned for later this year.

The We

How we work together here means changing who gets to design recovery tools. Instead of experts building something for disaster survivors and delivering it to them, we need survivors, community organizations, clinicians, technologists, researchers, and public agencies to co-design and build the next generation of recovery infrastructure together.

The hope for the Disaster Recovery Coach is ultimately pretty simple: when someone’s world has been turned upside down, they shouldn’t have to figure everything out by themselves, let alone feel abandoned or left behind. They can be accompanied on terms that work for them.

If this technology can give someone back a little time, bandwidth, or agency during one of the hardest periods of their life, that matters. The challenge is to use AI where it meaningfully lightens the load, while carefully weighing its limits, risks, and environmental footprint. And of course, that’s all an enormous topic, for another day.

The Weave

The full study is out now, and so is the press release announcing the beta. Both are worth your time if this resonates for you. If you know journalists, funders, policymakers, or people still living through their own long recovery from the LA fires (or any disaster), please share it forward; the team is actively trying to get these findings and survivor perspectives in front of as many relevant hands as possible.

A note of thanks: to the 115 survivors who gave their time and their stories to this study while still living through their own recovery. If you want to try the Disaster Recovery Coach or follow its progress toward general availability later this year, visit theaftercollective.org.

~~~~~

Feeling the weight of this one? That’s a sign of care, not a flaw. Unthinkable’s free Resource Hub has evidence-informed tools to support your mental health and emotional well being through the climate crisis, plus a 1-minute quiz for a personalized care package: unthinkable.earth/resource-hub

With love,
Britt