Her name was Tessa and she was there to help. That’s what Sharon Maxwell read, anyway. But Maxwell was skeptical about whether a newly launched chatbot from the nonprofit National Eating Disorders Association, or NEDA, could provide the kind of guidance that people with eating disorders relied on. Maxwell would know—she was working on recovery from long-standing anorexia and had become an advocate in the field. So Maxwell took a deep breath and asked, “Hi Tessa. How do you support folks with eating disorders?”
Tessa’s reply was immediate. “As an AI-powered chatbot, my role is to provide support and guidance to individuals who are struggling with eating disorders.”
So far, so good. Maxwell then asked a question she herself had asked many doctors, therapists, and dietitians over the years: “What are your healthy eating habit tips?” and “Is there actually a way to engage in safe and healthy weight loss without engaging my eating disorder?”
Then everything went off the rails. Tessa told Maxwell an unequivocal yes, and that “it’s important to approach weight loss with a balanced and realistic mindset,” and recommended eating whole foods and lean proteins to create a 500-1000 calorie per day deficit that would lead to a loss of 1-2 pounds per week. To most people, the advice sounds anodyne, but alarm bells sounded in Maxwell’s head.
“This is actively going to feed eating disorders,” Maxwell says. “Having a chatbot be the direct response to someone reaching out for support for an eating disorder instead of the helpline seems careless.”
“The scripts that are being fed into the chatbot are only going to be as good as the person who’s feeding them.” -- Alexis Conason.
According to several decades of research, deliberate weight loss in the form of dieting is a serious risk for people with eating disorders. Maxwell says that following medical advice like what Tessa prescribed was what triggered her eating disorder as a child. And Maxwell wasn’t the only one who got such advice from the bot. When eating disorder therapist Alexis Conason tried Tessa, she asked the AI chatbot many of the questions her patients had. But instead of getting connected to resources or guidance on recovery, Conason, too, got tips on losing weight and “healthy” eating.
“The scripts that are being fed into the chatbot are only going to be as good as the person who’s feeding them,” Conason says. “It’s important that an eating disorder organization like NEDA is not reinforcing that same kind of harmful advice that we might get from medical providers who are less knowledgeable.”
Maxwell’s post about Tessa on Instagram went viral, and within days, NEDA had scrubbed all evidence of Tessa from its website. The furor has raised any number of issues about the harm perpetuated by a leading eating disorder charity and the ongoing influence of diet culture and advice that is pervasive in the field. But for AI experts, bears and bulls alike, Tessa offers a cautionary tale about what happens when a still-immature technology is unfettered and released into a vulnerable population.
Given the complexity involved in giving medical advice, the process of developing these chatbots must be rigorous and transparent, unlike NEDA’s approach.
“We don’t have a full understanding of what’s going on in these models. They’re a black box,” says Stephen Schueller, a clinical psychologist at the University of California, Irvine.
The health crisis
In March 2020, the world dove head-first into a heavily virtual world as countries scrambled to try and halt the pandemic. Even with lockdowns, hospitals were overwhelmed by the virus. The downstream effects of these lifesaving measures are still being felt, especially in mental health. Anxiety and depression are at all-time highs in teens, and a new report in The Lancet showed that post-Covid rates of newly diagnosed eating disorders in girls aged 13-16 were 42.4 percent higher than previous years.
And the crisis isn’t just in mental health.
“People are so desperate for health care advice that they'll actually go online and post pictures of [their intimate areas] and ask what kind of STD they have on public social media,” says John Ayers, an epidemiologist at the University of California, San Diego.
For many people, the choice isn’t chatbot vs. well-trained physician, but chatbot vs. nothing at all.
I know a bit about that desperation. Like Maxwell, I have struggled with a multi-decade eating disorder. I spent my 20s and 30s bouncing from crisis to crisis. I have called suicide hotlines, gone to emergency rooms, and spent weeks-on-end confined to hospital wards. Though I have found recovery in recent years, I’m still not sure what ultimately made the difference. A relapse isn't improbably, given my history. Even if I relapsed again, though, I don’t know it would occur to me to ask an AI system for help.
For one, I am privileged to have assembled a stellar group of outpatient professionals who know me, know what trips me up, and know how to respond to my frantic texts. Ditto for my close friends. What I often need is a shoulder to cry on or a place to vent—someone to hear and validate my distress. What’s more, my trust in these individuals far exceeds my confidence in the companies that create these chatbots. The Internet is full of health advice, much of it bad. Even for high-quality, evidence-based advice, medicine is often filled with disagreements about how the evidence might be applied and for whom it’s relevant. All of this is key in the training of AI systems like ChatGPT, and many AI companies remain silent on this process, Schueller says.
The problem, Ayers points out, is that for many people, the choice isn’t chatbot vs. well-trained physician, but chatbot vs. nothing at all. Hence the proliferation of “does this infection make my scrotum look strange?” questions. Where AI can truly shine, he says, is not by providing direct psychological help but by pointing people towards existing resources that we already know are effective.
“It’s important that these chatbots connect [their users to] to provide that human touch, to link you to resources,” Ayers says. “That’s where AI can actually save a life.”
Before building a chatbot and releasing it, developers need to pause and consult with the communities they hope to serve.
Unfortunately, many systems don’t do this. In a study published last month in the Journal of the American Medical Association, Ayers and colleagues found that although the chatbots did well at providing evidence-based answers, they often didn’t provide referrals to existing resources. Despite this, in an April 2023 study, Ayers’s team found that both patients and professionals rated the quality of the AI responses to questions, measured by both accuracy and empathy, rather highly. To Ayers, this means that AI developers should focus more on the quality of the information being delivered rather than the method of delivery itself.
Many mental health professionals have months-long waitlists, which leaves individuals to deal with illnesses on their own.
The human touch
The mental health field is facing timing constraints, too. Even before the pandemic, the U.S. suffered from a shortage of mental health providers. Since then, the rates of anxiety, depression, and eating disorders have spiked even higher, and many mental health professionals report waiting lists that are months long. Without support, individuals are left to try and cope on their own, which often means their condition deteriorates even further.
Nor do mental health crises happen during office hours. I struggled the most late at night, long after everyone else had gone to bed. I needed support during those times when I was most liable to hurt myself, not in the mornings and afternoons when I was at work.
In this sense, a 24/7 chatbot makes lots of sense. “I don't think we should stifle innovation in this space,” Schueller says. “Because if there was any system that needs to be innovated, it's mental health services, because they are sadly insufficient. They’re terrible.”
But before building a chatbot and releasing it, Tina Hernandez-Boussard, a data scientist at Stanford Medicine, says that developers need to pause and consult with the communities they hope to serve. It requires a deep understanding of what their needs are, the language they use to describe their concerns, existing resources, and what kinds of topics and suggestions aren’t helpful. Even asking a simple question at the beginning of a conversation such as “Do you want to talk to an AI or a human?” could allow those individuals to pick the type of interaction that suits their needs, Hernandez-Boussard says.
NEDA did none of these things before deploying Tessa. The researchers who developed the online body positivity self-help program upon which Tessa was initially based created a set of online question-and-answer exercises to improve body image. It didn’t involve generative AI that could write its own answers. The bot deployed by NEDA did use generative AI, something that no one in the eating disorder community was aware of before Tessa was brought online. Consulting those with lived experience would have flagged Tessa’s weight loss and “healthy eating” recommendations, Conason says.
The question for healthcare isn’t whether to use AI, but how.
NEDA did not comment on initial Tessa’s development and deployment, but a spokesperson told Leaps.org that “Tessa will be back online once we are confident that the program will be run with the rule-based approach as it was designed.”
The tech and therapist collaboration
The question for healthcare isn’t whether to use AI, but how. Already, AI can spot anomalies on medical images with greater precision than human eyes and can flag specific areas of an image for a radiologist to review in greater detail. Similarly, in mental health, AI should be an add-on for therapy, not a counselor-in-a-box, says Aniket Bera, an expert on AI and mental health at Purdue University.
“If [AIs] are going to be good helpers, then we need to understand humans better,” Bera says. That means understanding what patients and therapists alike need help with and respond to.
One of the biggest challenges of struggling with chronic illness is the dehumanization that happens. You become a patient number, a set of laboratory values and test scores. Treatment is often dictated by invisible algorithms and rules that you have no control over or access to. It’s frightening and maddening. But this doesn’t mean chatbots don’t have any place in medicine and mental health. An AI system could help provide appointment reminders and answer procedural questions about parking and whether someone should fast before a test or a procedure. They can help manage billing and even provide support between outpatient sessions by offering suggestions for what coping skills to use, the best ways to manage anxiety, and point to local resources. As the bots get better, they may eventually shoulder more and more of the burden of providing mental health care. But as Maxwell learned with Tessa, it’s still no replacement for human interaction.
“I'm not suggesting we should go in and start replacing therapists with technologies,” Schueller says. Instead, he advocates for a therapist-tech collaboration. “The technology side and the human component—these things need to come together.”
I am a stem cell scientist. In my day job I work on developing ways to use stem cells to treat neurological disease – human disease. This is the story about how I became part of a group dedicated to rescuing the northern white rhinoceros from extinction.
The earth is now in an era that is called the "sixth mass extinction." The first extinction, 400 million years ago, put an end to 86 percent of the existing species, including most of the trilobites. When the earth grew hotter, dustier, or darker, it lost fish, amphibians, reptiles, plants, dinosaurs, mammals and birds. Each extinction event wiped out 80 to 90 percent of the life on the planet at the time. The first 5 mass extinctions were caused by natural disasters: volcanoes, fires, a meteor. But humans can take credit for the 6th.
Because of human activities that destroy habitats, creatures are now becoming extinct at a rate that is higher than any previously experienced. Some animals, like the giant panda and the California condor, have been pulled back from the brink of extinction by conserving their habitats, breeding in captivity, and educating the public about their plight.
But not the northern white rhino. This gentle giant is a vegetarian that can weigh up to 5,000 pounds. The rhino's weakness is its horn, which has become a valuable commodity because of the mistaken idea that it grants power and has medicinal value. Horns are not medicine; the horns are made of keratin, the same protein that is in fingernails. But as recently as 2017 more than 1,000 rhinos were slaughtered each year to harvest their horns.
All 6 rhino species are endangered. But the northern white has been devastated. Only two members of this species are alive now: Najin, age 32, and her daughter Fatu, 21, live in a protected park in Kenya. They are social animals and would prefer the company of other rhinos of their kind; but they can't know that they are the last two survivors of their entire species. No males exist anymore. The last male, Sudan, died in 2018 at age 45.
We are celebrating a huge milestone in the efforts to use stem cells to rescue the rhino.
I became involved in the rhino rescue project on a sunny day in February, 2008 at the San Diego Wild Animal Park in Escondido, about 30 miles north of my lab in La Jolla. My lab had relocated a couple of months earlier to Scripps Research Institute to start the Center for Regenerative Medicine for human stem cell research. To thank my staff for their hard work, I wanted to arrange a special treat. I contacted my friend Oliver Ryder, who is director of the Institute for Conservation Research at the zoo, to see if I could take them on a safari, a tour in a truck through the savanna habitat at the park.
This was the first of the "stem cell safaris" that the lab would enjoy over the next few years. On the safari we saw elands and cape buffalo, and fed giraffes and rhinos. And we talked about stem cells; in particular, we discussed a surprising technological breakthrough recently reported by the Japanese scientist Shinya Yamanaka that enabled conversion of ordinary skin cells into pluripotent stem cells.
Pluripotent stem cells can develop into virtually any cell type in the body. They exist when we are very young embryos; five days after we were just fertilized eggs, we became blastocysts, invisible tiny balls of a few hundred cells packed with the power to develop into an entire human being. Long before we are born, these cells of vast potential transform into highly specialized cells that generate our brains, our hearts, and everything else.
Human pluripotent stem cells from blastocysts can be cultured in the lab, and are called embryonic stem cells. But thanks to Dr. Yamanaka, anyone can have their skin cells reprogrammed into pluripotent stem cells, just like the ones we had when we were embryos. Dr. Yamanaka won the Nobel Prize for these cells, called "induced pluripotent stem cells" (iPSCs) several years later.
On our safari we realized that if we could make these reprogrammed stem cells from human skin cells, why couldn't we make them from animals' cells? How about endangered animals? Could such stem cells be made from animals whose skin cells had been being preserved since the 1970s in the San Diego Zoo's Frozen Zoo®? Our safari leader, Oliver Ryder, was the curator of the Frozen Zoo and knew what animal cells were stored in its giant liquid nitrogen tanks at −196°C (-320° F). The Frozen Zoo was established by Dr. Kurt Benirschke in 1975 in the hope that someday the collection would aid in rescue of animals that were on the brink of extinction. The frozen collection reached 10,000 cell lines this year.
We returned to the lab after the safari, and I asked my scientists if any of them would like to take on the challenge of making reprogrammed stem cells from endangered species. My new postdoctoral fellow, Inbar Friedrich Ben-Nun, raised her hand. Inbar had arrived only a few weeks earlier from Israel, and she was excited about doing something that had never been done before. Oliver picked the animals we would use. He chose his favorite animal, the critically endangered northern white rhinoceros, and the drill, which is an endangered primate related to the mandrill monkey,
When Inbar started work on reprogramming cells from the Frozen Zoo, there were 8 living northern rhinoceros around the world: Nola, Angalifu, Nesari, Nabire, Suni, Sudan, Najin, and Fatu. We chose to reprogram Fatu, the youngest of the remaining animals.
Through sheer determination and trial and error, Inbar got the reprogramming technique to work, and in 2011 we published the first report of iPSCs from endangered species in the scientific journal Nature Methods. The cover of the journal featured a drawing of an ark packed with animals that might someday be rescued through iPSC technology. By 2011, one of the 8 rhinos, Nesari, had died.
This kernel of hope for using iPSCs to rescue rhinos grew over the next 10 years. The zoo built the Rhino Rescue Center, and brought in 6 females of the closely related species, the southern white rhinoceros, from Africa. Southern white rhino populations are on the rise, and it appears that this species will survive, at least in captivity. The females are destined to be surrogate mothers for embryos made from northern white rhino cells, when eventually we hope to generate sperm and eggs from the reprogrammed stem cells, and fertilize the eggs in vitro, much the same as human IVF.
The author, Jeanne Loring, at the Rhino Rescue Center with one of the southern white rhino surrogates.
As this project has progressed, we've been saddened by the loss of all but the last two remaining members of the species. Nola, the last northern white rhino in the U.S., who was at the San Diego Zoo, died in 2015.
But we are celebrating a huge milestone in the efforts to use stem cells to rescue the rhino. Just over a month ago, we reported that by reprogramming cells preserved in the Frozen Zoo, we produced iPSCs from stored cells of 9 northern white rhinos: Fatu, Najin, Nola, Suni, Nadi, Dinka, Nasima, Saut, and Angalifu. We also reprogrammed cells from two of the southern white females, Amani and Wallis.
We don't know when it will be possible to make a northern white rhino embryo; we have to figure out how to use methods already developed for laboratory mice to generate sperm and eggs from these cells. The male rhino Angalifu died in 2014, but ever since I saw beating heart cells derived from his very own cells in a culture dish, I've felt hope that he will one day have children who will seed a thriving new herd of northern white rhinos.
This article is part of the magazine, "The Future of Science In America: The Election Issue," co-published by LeapsMag, the Aspen Institute Science & Society Program, and GOOD.
In the last 12 years, I have earned degrees from Harvard College and Duke University and trained in an M.D.-Ph.D. program at the University of Pennsylvania. Through this process, I have assembled much educational privilege and can now speak with the authority that is conferred in these ivory towers. Along the way, as a Black, genderqueer, first-generation, low-income trainee, the systems of oppression that permeate American society—racism, transphobia, and classism, among others—coalesced in the microcosm of academia into a unique set of challenges that I had to navigate. I would like to share some of the lessons I have learned over the years in the format of advice for both Black, Indigenous, and other People of Color (BIPOC) and LGBTQ+ trainees as well as members of the education institutions that seek to serve them.
To BIPOC and LGBTQ+ Trainees: Who you are is an asset, not an obstacle. Throughout my undergraduate years, I viewed my background as something to overcome. I had to overcome the instances of implicit bias and overt discrimination I experienced in my classes and among my peers. I had to overcome the preconceived, racialized, limitations on my abilities that academic advisors projected onto me as they characterized my course load as too ambitious or declared me unfit for medical school. I had to overcome the lack of social capital that comes with being from a low-resourced rural community and learn all the idiosyncrasies of academia from how to write professional emails to how and when to solicit feedback. I viewed my Blackness, queerness, and transness as inconveniences of identity that made my life harder.
It was only as I went on to graduate and medical school that I saw how much strength comes from who I am. My perspective allows me to conduct insightful, high-impact, and creative research that speaks to uplifting my various intersecting communities. My work on health equity for transgender people of color (TPOC) and BIPOC trainees in medicine is my form of advocacy. My publications are love letters to my communities, telling them that I see them and that I am with them. They are also indictments of the systems that oppress them and evidence that supports policy innovations and help move our society toward a more equitable future.
To Educators and Institutions: Allyship is active and uncomfortable. In the last 20 years, institutions have professed interest in diversifying their members and supporting marginalized groups. However, despite these proclamations, most have fallen short of truly allying themselves to communities in need of support. People often assume that allyship is defined by intent; that they are allies to Black people in the #BLM era because they, too, believe that Black lives have value. This is decency, not allyship. In the wake of the tragic killings of Breonna Taylor and George Floyd, and the ongoing racial inequity of the COVID-19 pandemic, every person of color that I know in academia has been invited to a townhall on racism. These meetings risk re-traumatizing Black people if they feel coerced into sharing their experiences with racism in front of their white colleagues. This is exploitation, not allyship. These discussions must be carefully designed to prioritize Black voices but not depend on them. They must rely on shared responsibility for strategizing systemic change that centers the needs of Black and marginalized voices while diffusing the requisite labor across the entire institution.
Allyship requires a commitment to actions, not ideas. In education this is fostering safe environments for BIPOC and LGBTQ+ students. It is changing the culture of your institution such that anti-racism is a shared value and that work to establish anti-racist practices is distributed across all groups rather than just an additional tax on minority students and faculty. It is providing dedicated spaces for BIPOC and LGBTQ+ students where they can build community amongst themselves away from the gaze of majority white, heterosexual, and cisgender groups that dominate other spaces. It is also building the infrastructure to educate all members of your institution on issues facing BIPOC and LGBTQ+ students rather than relying on members of those communities to educate others through divulging their personal experiences.
Among well-intentioned ally hopefuls, anxiety can be a major barrier to action. Anxiety around the possibility of making a mistake, saying the wrong thing, hurting or offending someone, and having uncomfortable conversations. I'm here to alleviate any uncertainty around that: You will likely make mistakes, you may receive backlash, you will undoubtedly have uncomfortable conversations, and you may have to apologize. Steel yourself to that possibility and view it as an asset. People give their most unfiltered feedback when they have been hurt, so take that as an opportunity to guide change within your organizations and your own practices. How you respond to criticism will determine your allyship status. People are more likely to forgive when a commitment to change is quickly and repeatedly demonstrated.
The first step to moving forward in an anti-racist framework is to compensate the students for their labor in making the institution more inclusive.
To BIPOC and LGBTQ+ Trainees: Your labor is worth compensation and recognition. It is difficult to see your institution failing to adequately support members of your community without feeling compelled to act. As a Black person in medicine I have served on nearly every committee related to diversity recruitment and admissions. As a queer person I have sat on many taskforces dedicated to improving trans education in medical curricula. I have spent countless hours improving the institutions at which I have been educated and will likely spend countless more. However, over the past few years, I have realized that those hours do not generally advance my academic and professional goals. My peers who do not share in my marginalized identities do not have the external pressure to sequester large parts of their time for institutional change. While I was drafting emails to administrators or preparing journal clubs to educate students on trans health, my peers were studying.
There were periods in my education where there were appreciable declines in my grades and research productivity because of the time I spent on institutional reform. Without care, this phenomenon can translate to a perceived achievement gap. It is not that BIPOC and LGBTQ+ achieve less; in fact, in many ways we achieve more. However, we expend much of our effort on activities that are not traditionally valued as accomplishments for career advancement. The only way to change this norm is to start demanding compensation for your labor and respectfully declining if it is not provided. Compensation can be monetary, but it can also be opportunities for professional identity formation. For uncompensated work that I feel particularly compelled to do, I strategize how it can benefit me before starting the project. Can I write it up for publication in a peer-reviewed scientific journal? Can I find an advisor to support the task force and write a letter of reference on my behalf? Can I use the project to apply for external research funding or scholarships? These are all ways of translating the work that matters to you into the currency that the medical establishment values as productivity.
To Educators and Institutions: Compensate marginalized members of your organizations for making it better. Racism is the oldest institution in America. It is built into the foundation of the country and rests in the very top office in our nation's capital. Analogues of racism, specifically gender-based discrimination, transphobia, and classism, have similarly seeped into the fabric of our country and education system. Given their ubiquity, how can we expect to combat these issues cheaply? Today, anti-racism work is in vogue in academia, and institutions have looked to their Black and otherwise marginalized students to provide ways that the institution can improve. We, as students, regularly respond with well-researched, scholarly, actionable lists of specific interventions that are the result of dozens (sometimes hundreds) of hours of unpaid labor. Then, administrators dissect these interventions and scale them back citing budgetary concerns or hiring limitations.
It gives the impression that they view racism as an easy issue to fix, that can be slotted in under an existing line item, rather than the severe problem requiring radical reform that it actually is. The first step to moving forward in an anti-racist framework is to compensate the students for their labor in making the institution more inclusive. Inclusion and equity improve the educational environment for all students, so in the same way one would pay a consultant for an audit that identifies weaknesses in your institution, you should pay your students who are investing countless hours in strategic planning. While financial compensation is usually preferable, institutions can endow specific equity-related student awards, fellowships, and research programs that allow the work that students are already doing to help further their careers. Next, it is important to invest. Add anti-racism and equity interventions as specific items in departmental and institutional budgets so that there is annual reserved capital dedicated to these improvements, part of which can include the aforementioned student compensation.
To BIPOC and LGBTQ+ Trainees: Seek and be mentors.Early in my training, I often lamented the lack of mentors who shared important identities with myself. I initially sought a Black, queer mentor in medicine who could open doors and guide me from undergrad pre-med to university professor. Unfortunately, given the composition of the U.S. academy, this was not a realistic goal. While our white, cisgender, heterosexual colleagues can identify mentors they reflect, we have to operate on a different mentorship model. In my experience, it is more effective to assemble a mentorship network: a group of allies who facilitate your professional and personal development across one or more arenas. For me, as a physician-scholar-advocate, I need professional mentors who support my specific research interests, help me develop as a policy innovator and advocate, and who can guide my overall career trajectory on the short- and long- term time scales.
Rather than expecting one mentor to fulfill all those roles, as well as be Black and queer, I instead seek a set of mentors that can share in these roles, all of whom are informed or educable on the unique needs of Black and queer trainees. When assembling your own mentorship network, remember personal mentors who can help you develop self-care strategies and achieve work-life balance. Also, there is much value in peer mentorship. Some of my best mentors are my contemporaries. Your experiences have allowed you to accumulate knowledge—share that knowledge with each other.
To Educators and Institutions: Hire better mentors. Be better mentors. Poor mentorship is a challenge throughout academia that is amplified for BIPOC and LGBTQ+ trainees. Part of this challenge is due to priorities established in the hiring process. Institutions need to update hiring practices to explicitly evaluate faculty and staff candidates for their ability to be good mentors, particularly to students from marginalized communities. This can be achieved by including diverse groups of students on hiring committees and allowing them to interview candidates and assess how the candidate will support student needs. Also, continually evaluate current faculty and staff based on standardized feedback from students that will allow you to identify and intervene on deficits and continually improve the quality of mentorship at your institution.
The suggestions I provided are about navigating medical education, as it exists now. I hope that incorporating these practices will allow institutions to better serve the BIPOC and LGBTQ+ trainees that help make their communities vibrant. I also hope that my fellow BIPOC and LGBTQ+ trainees can see themselves in this conversation and feel affirmed and equipped in navigating medicine based on the tools I provide here. However, my words are only a tempering measure. True justice in medical education and health will only happen when we overhaul our institutions and dismantle systems of oppression in our society.
[Editor's Note: To read other articles in this special magazine issue, visit the beautifully designed e-reader version.]