In today’s column, I examine the use of generative AI and large language models (LLMs) to provide immediate and potentially ongoing mental health support for people who have been psychologically impacted by a large-scale natural disaster.

This is an exceedingly leading-edge use of AI in the mental health realm and an aspect that I predict will become increasingly researched and ultimately accepted on a widespread basis. The advantage of AI is that people can readily access LLMs online and lean into AI-driven mental health guidance anywhere and 24/7. The cost is relatively low, especially when compared to accessing human therapists and other disaster support specialists. AI can be a valuable form of cognitive first aid and act as a means of getting people mentally back on their feet. Later, those who require further assistance can switch from AI and opt to make use of human-based mental health services.

Let’s talk about it.

This analysis of AI breakthroughs is part of my ongoing Forbes column coverage on the latest in AI, including identifying and explaining various impactful AI complexities (see the link here).

AI And Mental Health

As a quick background, I’ve been extensively covering and analyzing a myriad of facets regarding the advent of modern-era AI that produces mental health advice and performs AI-driven therapy. This rising use of AI has principally been spurred by the evolving advances and widespread adoption of generative AI. For an extensive listing of my well-over one hundred analyses and postings, see the link here and the link here.

There is little doubt that this is a rapidly developing field and that there are tremendous upsides to be had, but at the same time, regrettably, hidden risks and outright gotchas come into these endeavors, too. I frequently speak up about these pressing matters, including in an appearance on an episode of CBS’s 60 Minutes, see the link here.

Background On AI For Mental Health

I’d like to set the stage on how generative AI and large language models (LLMs) are typically used in an ad hoc way for mental health guidance. Millions upon millions of people are using generative AI as their ongoing advisor on mental health considerations (note that ChatGPT alone has over 900 million weekly active users, a notable proportion of which dip into mental health aspects, see my analysis at the link here). The top-ranked use of contemporary generative AI and LLMs is to consult with the AI on mental health facets; see my coverage at the link here.

This popular usage makes abundant sense. You can access most of the major generative AI systems for nearly free or at a super low cost, doing so anywhere and at any time. Thus, if you have any mental health qualms that you want to chat about, all you need to do is log in to AI and proceed forthwith on a 24/7 basis.

There are significant worries that AI can readily go off the rails or otherwise dispense unsuitable or even egregiously inappropriate mental health advice. Banner headlines in August of this year accompanied the lawsuit filed against OpenAI for their lack of AI safeguards when it came to providing cognitive advisement.

Despite claims by AI makers that they are gradually instituting AI safeguards, there are still a lot of downside risks of the AI doing untoward acts, such as insidiously helping users in co-creating delusions that can lead to self-harm. For my follow-on analysis of details about the OpenAI lawsuit and how AI can foster delusional thinking in humans, see my analysis at the link here. As noted, I have been earnestly predicting that eventually all of the major AI makers will be taken to the woodshed for their paucity of robust AI safeguards.

Today’s generic LLMs, such as ChatGPT, Claude, Gemini, Grok, and others, are not at all akin to the robust capabilities of human therapists. Meanwhile, specialized LLMs are being built to presumably attain similar qualities, but they are still primarily in the development and testing stages. See my coverage at the link here.

The Mental Impact Of Natural Disasters

Let’s focus on mental health impacts for those involved in or otherwise somehow connected to an association with a large-scale disaster. I will then tie this to a myriad of ways that people can get mental help, including the advent of leveraging AI-based mental health capabilities.

First, according to online stats by Statista, the United States is home to extremely diverse climates and, therefore, equally diverse natural disasters: “In the United States, 81 natural disasters occurred in 2024. Of these, 49 of them were severe convective storms, nine were caused by wildfire, heat waves, and drought, and another 12 were due to floods and flash floods. The few remaining were caused by winter storms, cold waves, and tropical cyclones. In addition to being extreme natural events, are also characterized by significant economic damage and/or loss of life.”

The gist is that a natural disaster can be greatly destructive in many ways. Some of the means are perhaps obvious, such as the regrettable loss of life and the massive destruction of property. What might not be quite so obvious is the mental toll of natural disasters.

Cascading Impacts

Think about the cascading aspects of mental health consequences:

• Level 1:Direct mental anguish of those who have been immersed in a natural disaster.
• Level 2:Mental suffering for those who know a loved one or other associate who was directly impacted and might be experiencing mental anguish (i.e., concerning those in Level 1).
• Level 3:Mental complications for those who acted during the natural disaster to help save lives and protect property, and witnessed the natural disaster first-hand or its aftermath.
• Level 4:Mental considerations for the loved ones and other associates of those people who had acted to assist during the natural disaster (i.e., those in Level 3).
• Level 5 and beyond:Others that, in some fashion, are mentally thrown off balance because of the natural disaster, even if not directly or immediately involved per se.

It is a daunting task to marshal mental health resources for a large-scale natural disaster. There could be hundreds, thousands, or possibly millions of people who require some form of mental health therapy. Where would all those needed mental health specialists come from? What would be the cost of engaging them in their efforts? How would those needing access to the mental health therapists arrange to do so? Etc.

It is a tremendously complex logistics problem, plus the enormous cost can be prohibitive.

Research On Disasters And Mental Health Services

In a posted analysis of mental health services and natural disasters, a study entitled “Mental Health Service Utilization After Natural Disasters: A Systematic Review” by Cate F. Woods, Virginia V. W. McIntosh, Ben Beaglehole, Caroline Bell, Psychiatric Online, January 8, 2026, made these salient points (excerpts):

• “Natural disasters are associated with new-onset mental disorders and distress and may exacerbate preexisting mental health issues.”
• “To date, no systematic review has examined rates and correlates of mental health service (MHS) use among adults exposed to natural disasters. This study aimed to synthesize findings examining MHS use among adults exposed to natural disasters.”
• “Forty-one articles (representing 39 studies) were identified across 18 countries.”
• “Rates of post-disaster MHS use are likely to depend on disaster-specific contextual factors.”
• “Among participants reporting post-disaster MHS use, nonspecialist primary health care services were utilized most frequently. Mental health recovery efforts may benefit from prioritizing nonspecialist primary health care services and utilizing outreach approaches targeted toward groups with high degrees of exposure severity.”

I’d like to emphasize that, according to the study, there isn’t an extensive amount of research on the topic of mental health services that arise for coping with natural disasters. It is a realm that is woefully understudied. Thus, we don’t have much empirical data to readily understand the nature and extent of natural disasters as tied to the engagement of mental health services.

Of the sparse data available, it appears that post-disaster use of mental health services tends to invoke nonspecialist primary health care services. This makes logical sense. A person goes to see their primary care provider and at that time reveals or the service ascertains that a mental health matter is at play. The concern is that primary care might brush off the matter or fail to even look for any mental health issues. It isn’t necessarily in their bailiwick.

Perhaps other means could shore up this conundrum.

AI As Psychological First Aid

The use of generative AI is a timely and useful tool for providing mental health assistance regarding post-disaster circumstances. AI can be a type of psychological first aid.

As noted earlier, AI is readily available online. It can be accessed from just about anywhere. It can be utilized 24/7. The key is that AI is an “at scale” option to handle a circumstance that is undoubtedly a large-scale consideration. We are matching the magnitude of the problem with a solution that works on an equally large scale.

Critics might insist that ordinary generic AI, such as the popular ChatGPT or GPT-5, isn’t particularly suitable for providing mental health guidance in the specific context of natural disasters. In that sense, the idea is that there would be customized AI that is tuned to the particulars underlying mental guidance regarding natural disasters. The use of generic AI might be undertaken, but if this is to be sanctioned by a governmental entity, the odds are that a customized approach would be more appropriate.

Important Questions To Consider

That being said, you can bet your bottom dollar that people are going to turn to generic AI anyway. Think of it this way. A person is perhaps already using ChatGPT or some other popular LLM on a daily or weekly basis to give them insights on how to fix their car or make chocolate chip cookies. When a natural disaster has occurred, and the person believes they need some mental help, they are almost surely going to return to the AI that they already know and frequently use.

Here are some crucial questions to ponder:

• Should AI makers ensure that their generic AI is suitably ready to help people who are seeking mental health support after a natural disaster?
• If AI makers are not doing so, should policymakers and lawmakers require that they do so?
• What responsibility does an AI maker have if their generic AI goes awry and provides untoward advice to a person seeking mental health support after a natural disaster?
• And so on.

Ways That AI Can Be Helpful

Suppose we don’t want AI to act like a therapist in these situations. It could be that we collectively decided that AI performing therapeutic duties is construed as a bridge too far. Okay, in that case, there are various adjacent ways that AI can still be helpful for the mental wherewithal of those in these predicaments.

Consider these possibilities:

• Stabilization:AI provides grounding exercises such as breathing guidance, meditation, and the like.
• Psychoeducation:AI explains how stress and trauma are exhibited and provides an educational perspective on mental issues that could be taking place.
• Emotional release:AI undertakes active listening of the person, reflects on what they have said, summarizes, and goes no further (avoiding providing any type of diagnosis).
• Cognitive off-ramping:AI serves to interrupt catastrophic thinking that a person might be bogged down in, nudging them to slow down and reassess the circumstances.

The crux of those approaches is that they are in a low-risk category of what the AI can help with. By sidestepping the therapy aspects, the AI is pretty much a handy shoulder to cry on. It isn’t overtly diagnosing or ascertaining mental health conditions. You could make the case that the AI is acting within the realm of non-clinicians, such as disaster volunteers.

Triage And Referral Via AI

There is more that the AI can do, while still staying out of the therapeutic domain.

The AI could attempt to determine whether someone is having a potentially disconcerting reaction to the natural disaster. Red flags could be detected. This might be done by asking standardized screening questions, such as those found on the psychological instruments PHQ-9 or GAD-7.

If a person appeared to be especially out of sorts, the AI could refer the person to a human therapist. Maybe special crisis counseling resources have been set up for the natural disaster, and the AI could point the person to go there. The AI acts as a triage and referral mechanism.

One objection to the AI doing this is that the AI might generate false positives or false negatives. Let’s examine those downsides.

The Worrisome Downsides

Suppose the AI doesn’t refer someone to human assistance, but the AI should have done so. Though that is troubling, the AI could be shaped to always inform a user right away about available human post-disaster resources. The AI wouldn’t hold back that info. It would present it at the get-go and merely provide a reminder if the detection suggested that something more serious is going on.

Next, what if the AI falsely refers a person to human assistance even though they don’t actually need it? That’s a relatively minor downside. The odds are that lots of people are already seeking out those resources.

Those resources would presumably already have a screening process in place. If the AI referred someone, and the screening said they don’t belong, that’s not something that excessively taxes those services. You could make a case that by having people go through the AI first, the volume of those having to go through screening is going to come down immensely. The AI is saving those precious resources from overconsumption.

AI Safety And AI Governance

A slew of notable design aspects should be considered when it comes to AI safety and the governance of AI that is serving this capacity.

The matter of privacy and data protection is at the forefront of this. A person using AI in this manner is likely highly vulnerable and going to divulge intense personal thoughts to the AI. How will that data be kept private? How long will the data be retained? Does the AI adhere to relevant privacy laws? Etc.

AI safeguards need to be established. A person might say that their reaction to the natural disaster has led them to consider self-harm. The AI should not ignore such remarks. An immediate referral would likely be warranted, including explicit notification and not just simply urging the user to contact those outside resources.

On a governance basis, those are certainly vital considerations, along with the ability to designate AI that is officially working in this capacity and letting people know accordingly. The problem is that oftentimes, natural disaster victims and others are pursued by unscrupulous evildoers. They try to scam those people.

You can likely imagine how they might use AI in that same manner. They advertise that their AI is there for anyone who is suffering mental anguish from a natural disaster. Meanwhile, once someone uses the rigged AI, the scoundrels are actually stealing their private data and going to use the data for criminal purposes. The evildoers are riding on the coattails of the proper AI.

Lots of dynamic AI governance concerns require scrutiny and preparation.

Establishing Oversight

A designated public health agency or authority would be the proper place to establish oversight on this usage of AI. It’s a FEMA (Federal Emergency Management Agency) kind of milieu, encompassing affiliated federal, state, and local entities. You could readily add AI usage to the list of other tools being used for natural disasters, such as special hotlines and crisis text lines that are often put into use.

With AI, a bonus is that you could do after-action analyses. Since people are using a collective AI that contains their prompts and concerns connected to the natural disaster, it would be easy to do macroscopic analyses of trends and patterns. We could learn quite a bit about how people react mentally to various types of natural disasters. Perhaps mental issues differ from one kind of disaster to another, or have other subtle contextual dependencies that we didn’t realize are underway. AI would be a rich source of a public-wide database mindset that could be used to improve systematic responses to natural disasters on a massive societal scale.

The World Ahead

Let’s end with a big picture viewpoint.

It is incontrovertible that we are now amid a grandiose worldwide experiment when it comes to societal mental health. The experiment is that AI is being made available nationally and globally, which is either overtly or insidiously acting to provide mental health guidance of one kind or another. Doing so either at no cost or at a minimal cost. It is available anywhere and at any time, 24/7. We are all the guinea pigs in this wanton experiment.

The reason this is especially tough to consider is that AI has a dual-use effect. Just as AI can be detrimental to mental health, it can also be a huge bolstering force for mental health. A delicate tradeoff must be mindfully managed. Prevent or mitigate the downsides, and meanwhile make the upsides as widely and readily available as possible.

A final thought for now.

Catherine the Great famously made this remark: “I beg you take courage; the brave soul can mend even disaster.” The use of AI to mend people’s souls after a disaster might seem ironic that a machine could be of such help to humans, but the timeliness and aptness are worth a serious look-see. It often takes courage to try new innovations.