When States Step Up for Safety with Delegate Lily Qi

welcome to Mental Health Evolution, a podcast about what's changing in mental health and why it matters. I'm your host, Rachel Harrison, inviting you into honest conversations with with people from all perspectives in the field. Clinicians, tech founders, investors, insurance companies, and all the folks in between. Let's explore what's working, what's not, and what's next. All right everyone, welcome back to the Mental Health Evolution podcast where we are talking about how the landscape is quite quickly evolving in the mental health industry. As you probably know, we are coming to you from the State of Maryland, and today we are really excited to speak with Maryland State Delegate Lily Chee, who represents District 15 just south of us here in Montgomery County. Delegate Chi is currently seeking another term as a Maryland delegate. She works at the intersection of economic policy, consumer protection, and state systems design. In recent legislative sessions, she has focused on how Maryland responds to emerging technologies, including artificial intelligence, and how those technologies intersect with behavioral health systems and consumer safety. Her work sits within a broader national trend where states are increasingly stepping into mental health regulation, workforce challenges, and technology oversight in the absence of consistent federal standards. And as always, before we talk to our guest, we'd like to bring up a few relevant articles related to the topic today. These may be helpful for you if you know that there's some legislation out there, but you don't know exactly what we're talking about. And we'll definitely link all of these in the show notes for you so that you can have a great point of reference to go back to if you'd like to as well. The first article is a link to Maryland House Bill 883, which is which focused on the use of artificial intelligence in behavioral health care contexts. This was the bill that we were working to promote from our end in Maryland and what initially got us interested in talking to Delegate Chi today. Unfortunately, it did not make it through the full legislative process this session, but it's still an interesting example to look at in the context of how states are approaching AI regulation. Similar legislation has been passed in other states and in this episode we'll talk to Delegate Chi about plans to potentially reintroduce this or similar legislation in the upcoming session. And then the next article we have is talking about how states across the country are continuing to increase investments in mental health systems, including workforce development, school based services, crisis response, and behavioral health infrastructure. This reflects a broader recognition that mental health challenges are not only clinical issues, but system level issues that require sustained public investment at the state level, especially as federal funding priorities continue to shift. And this article is from NGA.org and lastly, we have a growing article number three is from AI2 and this is A growing number of states are beginning to pass legislation addressing the use of artificial intelligence in mental health contexts, including concerns about safety, accuracy, and whether AI tools should be allowed to simulate therapy or behavioral health services. Tennessee, for example, has moved to ban certain AI therapy chatbots, signaling a broader wave of state level regulation aimed at protecting consumers in the absence of federal standards standards. So you kind of see the framework of what we're talking about here. States are really trying to figure out the best way to balance the idea of what this technology can do to help us and also protect consumers in the process. So that is a lot of what delegate she is all about and I'm excited that she's here to talk with us. So welcome Delegate.

She thank you very much. Richer. It's a pleasure to join you today.

Thank you. So to start off, I'd love to just ground this for our listeners. You've worked on a wide range of issues in the Maryland General Assembly. What first got you interested in the intersection of artificial intelligence, consumer protection and behavioral health?

Great question. I have served on the Economic Matters Committee for my entire two terms and that means much of my work is at the intersection between consumer protection, business regulation and advancing Maryland's competitive industries of the future. As someone with an economic development background, I certainly will continue to champion our innovation industries, including AI and other related industries. So this is not about not wanting AI to develop in the state of Maryland or serve consumers. I personally also use chatbots for research, for verification, for ideas sometimes. Right. So definitely AI has a very important and increasingly important role to play in our everyday lives, has enormous amount of benefits that we can all enjoy, but at the same time because of our increasing relationship and dependency sometimes on AI usage for our everyday work and life, there's also a risk that is not talked about enough or. Or people are aware of, especially for the most vulnerable population, whether they are seniors or people who speak English as a second language, or our youth. So we all know, you know, since the pandemic, there was a huge surge of youth mental health crisis, and that surge coincided with the rapid evolution of AI. So we all knew something needed to be done. And as someone who pays attention to something like that, I was aware that the Trump administration just does not like the idea of the states regulating AI. In fact, there was an executive order specifically threatened to punish states that dare to regulate AI. So short of federal action, which I wished was the case, states have to be the backstop to protect our consumers and to have some baseline regulation and clarity about what is and is not acceptable behavior. So when this bill idea was brought to me, I thought it was perfect. This is a great timing for us to assume on this issue. Plus, I happen to serve on the right committee to advance that bill.

Yeah. All the things kind of aligned, it sounds like, Right? Yeah, yeah. And so I would love to hear your sort of story of how the bill got started, what happened, and maybe a little bit about what's next. I know that's kind of broad, but.

Yeah. So when the advocates brought this bill to my attention, I discussed it with my chief of staff and I said, you know, I'll be more than happy to do that. Of course, we have zoom calls, multiple zoom calls with different advocacy groups and bipartisan support to sign on as co sponsors, because I don't know if you're aware, but this session, the House Minority leader himself also introduced a similar bill because. So this is a bipartisan issue that affects everyone. Right. So I was very encouraged. The colleagues, my Republican colleagues offered to co sponsor my bill. He said, hey, we heard you are introducing a chatbot bill. Can we sign on to that? I said, perfect. You know, so I was very encouraged and got a long list of people to. Would sign up to testify at the bill hearing. You could see that online. And I was very encouraged that the bill, you know, got advanced all the way across the House with overwhelming bipartisan support. And I even presented it on the Senate side. Unfortunately, that got stalled. And I don't know if it was because there was confusion about the scope of it or was because Big Tech's aggressive lobby. Right. Because at the end of the day, as much as the tech companies say, you know, we don't want a 50 state patchwork, and I understand the bottom line, Is, you know, this is not the first industry that has to go through 50 state regulation. You know, banking, for example, is regulated at a federal level, but every state also has their own banking laws. Right. Insurance, you know, another major industry. Right. That affects every everyday people's lives is, is not heavily regulated at the federal level. So every insurance company will have property and casualty insurance, for example. Right. Will have to go through 50 different states rule books. So this is not new, a new territory for businesses to adapt to. Right. And so we all, you know, can leverage each other's examples and model policies to see what is working and whatever to make it work for tech companies. And frankly, at the end of the day, if the big techs were doing their jobs, right. So self regulating, we wouldn't be here talking about this. Right. I mean, at my hearing, there was a parent who came with this framed photo of his 12 year old boy who took his own life because of chatbot manipulation. Right. So that's kind of example we wanted to stop. There's nothing right now in Maryland that stops this kind of self harm or harming other people because of chatbot influence. So something has to be done. And because it got stalled, you know, next step is I will continue to build alliance, you know, continue to educate other legislators. And this year I think I would have the advantage of starting early, earlier. Right now that we have gone through the, this round and made some progress on the House side, I think we can be more focused in our efforts.

Yeah, yeah. And, and I think you're right. I too am not against AI in any capacity. Right. But as a, as a person in the mental health industry and being trained as a clinician, it is so alarming to think that there's this access out there for anyone. And, and there are lots of AI chatbots being marketed as a counselor until you look at the fine print and they say, no, we're not a therapist, we're Right. And so that's very confusing because all the marketing says yes, we're here to support you. And then it's like, but if you tell me something that's challenging or difficult, don't blame me, I'm not a therapist. Right. It's very. That's right. It's just risky. And when it's free also there's just no, no guardrails.

Right, Right. I don't know if in my response to your earlier question, I made it clear what the bill actually does is very simple, which is simple disclosure. Right. To say, you know, this AI system that you are Interacting with is an AI system. It's not certified. A licensed human professional at the back end giving you proper professional advice for mental health. Right. That's a simple disclosure. People need to know before they fall into the traps, you know, because the chatbots are designed to kind of coast your ego makes you feel good about the decisions and ideas you have. They keep encouraging you even if you're on the wrong path and say, oh, that's a great idea. Yeah, you know, you should do that. Right. Why not? You know, and it's designed to be so encouraging and stroke your ego and make you feel good about whatever ideas you have. So I think this is really a very high risk proposition we're looking at.

Yeah, I mean. Yeah, and I think that too. That's not. You're not asking too much. You are not. I think it's the state of Illinois that has said there can't be any AI therapists or chatbot therapists. Like there are definitely stronger lines being drawn in different states. So simply disclosing and then allowing people to have the information to make the decision then on their own what they choose to do with that seems like a pretty reasonable step.

Yeah, we thought it's very reasonable.

Yeah. What are your thoughts on some of the other state rulings? Not that I expect you to be an expert on all of those, but I'm curious if any of those informed or inspired your decision for what you put forward in the House bill.

Well, I am confident after this session, between now the next session, there will be more states that are stepping up introduce, or maybe this session they have already passed the virus. Maryland stalled. Right. AI chatbot bill. And by the way, the minority leaders bill also did not advance on the Senate side. So I think we need to focus on the Senate. Yes, it sounds that way. Yeah. I was encouraged that not only, you know, did Maryland House passed with overwhelming bipartisan support, but states, you know, other red and blue states, including those very tech driven states like California, New York. Right. Are leading the way, you know, and other states like the Newtown, Nevada, Illinois and all that have all stepped in. And you mentioned Tennessee. So this is, you know, going to sweep across the country as more and more Americans are waking up to the downside of unregulated AI and how invasive these systems can be in our lives. So I think I'm very optimistic and hopefully it's just a matter of technically how to do it. Right. And how not to let the tech lobbyists come back and say you're being too broad. I don't know, if you saw my pushback against their pushback, so to speak, and say, you know, this is too broad and blah, blah. I said, well, they come up with some solutions. Right, right. If you know the status quo is not perfect, then work with this. You know, make it work. Don't just come say, categorically oppose any regulation. As I said, you know, if you were good at self regulating, we wouldn't be here talking about this, you know, so work with us, give us ideas. If it's need, if it needs to be narrowed and tailored, then show us what works in other states.

Yeah. And I think especially in the mental and behavioral health space, it's just not a tool that can be broadly applied without some thought processes and guardrails and things like that. So I think it's been, yeah, maybe a little accidental, maybe a little shortsighted, maybe all those things. But I have talked on this podcast with some people who are developing different AI systems and a lot of them do acknowledge some of the errors, some of the hallucinations, the mistakes, the risks, you know, so that information is out there. Right. It's not, how can it be? It's pretty new.

Yeah, that's right. Yeah. There's human bias too in training the AI to make decisions.

Hmm. Yeah, absolutely. Absolutely. So when you and other legislators coordinate these approaches when it comes to build like this, when technology is moving faster than any one state can keep up with, how does that coordination and collaboration work? And is there anything that listeners can do to help support this?

Yeah, great questions. I'm looking forward to joining this year's national conference in Chicago, which is the national conference for state legislators. And with AI being so pervasive and important as an area of regulation, I'm really looking forward to joining these workshops and seminars to learn what other states are doing and how to perfect my bill to make it work to remove the impediments to passage, hopefully to three other path and how they build coalitions across different consumer groups and user groups and all of that. So I'm really looking forward to connecting with other state legislators, learning different ways to tackle this issue.

And is there any particular way that someone listening, like maybe they're in Maryland and this bill specifically, or maybe they're just in the United States and they're just interested in advocating for this. What would you recommend as next steps for that type of advocacy?

I think it's important to bring ideas to legislators because we are no subject matter experts in every aspect, in every area. Right. I came from very different professional background than AI regulation, for example. It is such a new area that everyone is learning and it's constantly evolving. So please bring new ideas to us and reach out to our offices. But also, if you feel passionate and strongly about certain things, build a coalition and don't just expect the hearing itself to deliver magic. The real work is behind the scenes. Right. Is reaching out to the committee members one by one. Committee chairs make sure both the House and Senate are on the same page. Work with us throughout the process. Don't just expect to show up and you know, speak your two minutes worth testimony and think people will get a message. You know, with thousands of bills introduced each year and only 90 day session, my advice to the advocates is start early, do your homework, reach out, build coalition. Right. Remove the potential questions that legislators may have on their minds about how this might work and all that. I mean, the more we do the homework, the more prepared we are to push for the passage of a bill like this.

Mm. And when you say create coalitions, what exactly do you mean by that?

I think a coalition could be built between the tech community, which also, at least you know, that's what they claim they care about users because that's what they sell to. Right. They cater to, to users. And I, and I don't doubt that they are doing a lot of work to stop the harm and all of that, but it's just not enough. Right. We have seen harm, we have seen real danger, you know, on our community, especially the vulnerable populations. So it needs regulation and we can never stop, you know, bad players from entering the market until you have regulation right in place. So, you know, I think community advocates, you know, parent groups or anyone who is concerned about the AI's harm, build coalition with other, you know, like consumer advocacy groups, for example, as well as the tech community. Right, right. And with the school systems, for example, or with the senior community, whatever places where you think people will get it and with the behavioral professionals or other licensed professionals that people often, or the kind of service providers where people often go to AI to seek help from advice. So I think there are multiple ways to do that and of course to build bridges and relationships with, with your own state representatives and legislators is always a good idea.

I think that's a great invitation coming from you, delegate G. Because I don't know that in my industry a lot of people talk about that or think about that necessarily. And it is such a changing industry and there is so much happening right now and that is a great. Yeah, just opening up lines of communication Giving our perspective, expertise, what's impacting us, all of those things. So I really appreciate you saying that. I don't think that that is often a route in mental health. We can be so focused on patient care that that's sort of all we focus on.

Right, right. You know, the immediate issue. Yeah.

And then these things come and they impact us and our clients, and we suddenly get worried, but then it's like, okay, well, wait, what do we do about this? And that's really how this podcast was born, is like, what do we do about all these changes? I don't know. So let's have the conversation. Let's see what people are talking about. Yeah, yeah. Zooming out a little bit. When you think about behavioral health more broadly, where do you see the biggest gaps right now that state policy is actually able to address in a meaningful way? That's a big question.

Yeah. I think in the stepping back, I think when you look at holistically, this space of AI regulation, I think we are still. The whole world is still relatively new. Right. Understanding AI and how it can be so sweeping in benefiting or harming our lives. And what should the human relationship be? How do we define healthy human relationship with an AI system? Right. So that we can reap the benefits without being harmed by the system, whether it's workforce disruption, whether it's, you know, AI taking over certain people's ability to think critically and do their homework and all of that. Right. But on the other hand, there's tremendous potential to. To improve the quality of life and the human experience in so many ways. Right. So I think this is still a new frontier, that when we take a step back, we should just look holistically. And then another issue is how Maryland wants to position ourselves in this whole AI revolution. Right. How inviting do we want to be in facilitating the development of AI companies and industries in our state as a competitive advantage? Right. But at the same time putting in proper guardrails so our community is not harmed or is aware of the potential harm of AI. And then another issue related to that is workforce shortage related to certain areas. In this case, that we're talking about behavioral health. You know, people turn to AI because just not enough resources around to help them with their mental and behavioral health. They think this is anonymous tool that makes them feel safe. That's why they turn to AI. Right. How do we create that kind of human environment where it's safe and accessible and hopefully also culturally competent so that we can meet people where they are and how we can invest in such workforce that is linguistically accessible, culturally competent, and hopefully also age appropriate and all of that. So there are a lot of big questions we have not answered. We are still grappling with. This is the kind of issues policymakers at the federal and the state level should be thinking about.

Yeah, that's pretty inspiring from my perspective. One of the things that I feel passing a bill offers is just that there is guidance out there for consumers.

It's.

It's a line in the sand that says be careful. Right. Some people may still choose to use it, but from the end of providing behavioral health, mental health services, you know, a lot of our clients are utilizing it alongside human therapy. So sometimes we're educating and talking about how do we use that interactively and effectively, but not dangerously. And the legislation to me gives some guidance to that, gives some support to that and really helps the work that we are able to do. So, yeah, I appreciate all that you have done in that vein. Anything as we wrap up here today that you would like to leave our listeners with as a last tidbit or something for them to be mindful of or encourage them or inspire them in any way?

Well, I just want to be encouraging by saying don't give up. You know, this is an important field and I will continue to be your partner. Please reach out to our office. If you want to be involved in advocating and pushing for the bill's passage hopefully next year, this is the time to have this kind of conversation. I welcome your ideas, your input, whatever. Hope to see you in and after and beyond. It is a partnership, an ongoing partnership.

Awesome. I appreciate that so much. Delegate G. Thank you so much for being here, talking through the process with us and yes, we look forward to next session as well.

Great. And maybe you want to share with your listeners my contact information so that.

Absolutely. We will include that in the show notes. If anybody wants to, to reach out, that would be great. And we'll have all those articles in the show notes and so people can also follow the progress and see what happened last session as well.

Great.

And for our listeners, thank you for diving in with us today. We will be back next week to talk more about the evolving landscape in the mental health field. Bye for now. Thank you.

Rachel.