Results from the First Field Study of AI for Legal Aid

Episode Notes

Transcript

Colleen Chien:

So it seemed like an opportunity to really try to think about if we want to deliver on the promise of AI being an equalizer, we need to actually deliberately engage in that question.

Announcer:

Equal access to justice is a core American value. In each episode of Talk Justice An LSC Podcast, we’ll explore ways to expand access to justice and illustrate why it is important to the legal community, business government, and the general public. Talk Justice is sponsored by the Leaders Council of the Legal Services Corporation.

Cat Moon:

Hello and welcome to Talk Justice. This is your host for today’s episode, Cat Moon, and today I am thrilled to speak with three people who are doing some really interesting and important work at the intersection of generative AI and legal aid. My three guests are Colleen Chien, who is a professor of law at Berkeley Law School, co-director of the Berkeley Center for Law and Technology, and founder of the Paper Prisons Initiative. Miriam Kim is also one of my guests and she, along with Colleen Coauthored a paper we’re going to be discussing today. Miriam is a partner at the law firm, Munger Tolls and Olson, and is a fellow of the Berkeley Center for Law and Technology. Our third guest today who worked with Colleen and Miriam on the project, which is described in the paper we’re going to discuss is Bréyon Austin, and she is pro bono manager for the Lawyers Committee for Civil Rights of the San Francisco Bay Area.

So Colleen and Miriam have done some research and written, I think a very important paper about their research. The name of the paper is Generative AI and Legal Aid Results from a Field Study and 100 Use Cases to Bridge the Access to Justice Gap. By the way, if you go to ssrn.com, you can do a search for the paper and pull it up and read it for yourself. I highly recommend the three of us are going to talk about the research described in the paper and its implications. Bréyon was actually a participant in the research study, and so she can talk about it from a very practical perspective. So without further ado, I would like to introduce Colleen, Miriam, and Bréyon, and yeah, have a fantastic conversation with them. Welcome to all three of you. Thank you so much for being here to have this conversation about what I think is really exciting research, and I want to dig right in and talk first about what inspired you all to engage in this particular research in the first place. How did this come to be?

Miriam Kim:

So how this all came to be is with the rise of chat GPT, I think there was a lot of interest in thinking about how legal technology could be used to increase access to justice. The state bar of California had already been working on those issues for some time, but they approached Berkeley Law School in the spring of 2023 because they had lost the ability to really pursue that as they had been in the past due to some funding and political battles that have arisen with respect to the state bar’s, work on those issues. And fortuitously, that was the same time that I was looking into the possibility of being a fellow at Berkeley during a sabbatical from my law practice. And so I was able to connect with the state bar of California and then connect with Colleen who was joining the faculty that year at Berkeley to think about how Berkeley might take a leadership role in trying to advance efforts to increase access to justice through technology,

Cat Moon:

Talk about timing.

Miriam Kim:

So I think it was really fortuitous timing and to add to just how fortuitous this is, Colleen and I were roommates in law school at Berkeley, and so we reunited at Berkeley, this really fortuitous moment with support from the state bar of California, from uc, Berkeley. And then we also went about getting support from the AI community and from the legal aid community to try to come together and collaborate and see what we might do to explore how generative AI could be used to increase access to justice.

Cat Moon:

That’s really an amazing story, fortuitous connections and how important all those connections can be to make something like this happen. So well I for 1:00 AM really glad. Colleen, is there anything you would like to add to that?

Colleen Chien:

Sure. So I think for me personally, as soon as I got to Berkeley Law, the focus on AI was so intensified in terms of thinking about the different regulatory issues, the business opportunities, the social implications of ai, but there seemed to be a sort of big gap between the equalizing promise of AI and delivery on that promise. I think on the one hand you have regulators who are worried about all the things that could go wrong, and then you have the market sort of going forward and trying to think about all the new applications that could be developed that could add value, but the idea that then eventually this can be an equalizing force is not something that has been sort of at the forefront. So it seemed like an opportunity to really try to think about if we want to deliver on the promise of AI being an equalizer, we need to actually deliberately engage in that question.

And then as a researcher and somebody empirical work, I thought that it’s a luxury to be able to really look at this issue in depth and to do sort of a rigorous pilot is to really try it and measure the outcomes using a randomized control trial. So the appeal of being able to do that made it a natural thing to proceed. And then of course the opportunity to work with Miriam was really significant. And I would just say again, at Berkeley Law, we had both participated a lot in the pro bono programs and I think we’ve both continue to do service since we’ve graduated. And so again, the law school seems like a good place in which to really bring all this together.

Cat Moon:

That really is an amazing story and very fortuitous, and we absolutely do need rigorous study of all of these things. And so I think you all are definitely breaking some ground here. I think you’re also providing really a structure and a format that hopefully other people can build upon. I think that’s important as well. So you did research with legal aid folks in legal aid, not just lawyers, but kind of a mix of people serving in different roles. And we’re very lucky we have one of those folks here with us. So Bréyonne, you chose to participate in this study and I’m curious what brought you to join the study?

Bréyon Austin:

Sure. So Miriam approached our team, my team at the Lawyers Committee for Civil Rights of the San Francisco Bay Area. This all actually happened while I was out on sabbatical. So I came back and they’re like, we’re doing this project. And I’m like, great, what is this project? So that was pretty exciting and I was trying to find a way to use the word fortuitous as well since we’ve been using that a lot, but maybe we can put that in there. So I’m really grateful to my team for getting us involved with this. And just a little background, we’re a nonprofit based in San Francisco, and as a part of that, we host a free drop in legal clinic in the Tenderloin District in San Francisco. And so through that clinic is how we were able to participate in this pilot. And when we discovered that the use of AI may have the potential to increase access to justice for our client communities, we were very excited to test it out. We’re always searching for tools that’ll help us better serve our clients and our communities, and we really also want it to be a part of sharing our experience with other legal professionals that are curious about all the ways in which AI can be utilized. So that’s how we got involved.

Cat Moon:

I think a significant aspect of this research, from my perspective as someone who is trying to facilitate understanding and knowledge among the legal community about generative AI and the benefits and risks you are making available, all of these fantastic use cases for folks and reporting on how helpful or not these different use cases were. And I think just sharing these examples from folks, really boots on the ground here are things we’ve done. So I can tell you I shared many of them yesterday with Legal Aid of Middle Tennessee in a conversation I had with their entire staff about using generative ai. And it was so incredibly valuable to me. I certainly could give them a laundry list of use cases from legal, but for them to hear from others in their shoes basically across the country, ways they’ve used it, I just thought that was incredibly valuable and incredibly powerful. And so really making all that transparent and available to other people. So I just want to thank all of you for that because I think that one simple thing alone is incredibly powerful and you share a database of a hundred use cases all rated in terms of the efficiency gain and how useful it was, whether it’s likely to be replicated. Miriam, I’m going to pass the virtual mic to you, and if you can just chime in, what are really some of your primary takeaways from the research?

Miriam Kim:

Some of my primary takeaways are just the importance of giving tools to those who are boots on the ground, who are in direct contact with people in need. We did that at the Lawyers Committee for Civil Rights by putting it in the hands of Bréyonon and other lawyers supervising the legal clinic because in California we still have unauthorized practice of law regulations that limit developers’ ability to create consumer facing legal technology tools. So at this stage of AI development, we thought it was important that we make sure that the legal aid lawyers who are in most contact with the pro se litigants or those who they are able to represent to figure out how generative AI can be used, and then we can then create a feedback loop back to the developers. So this is going to require funding to get the tools and the hands of those on the front lines, but also requires training and support. And Colleen came up with the brilliant idea she mentioned earlier, to do this as a randomized control trial. So we wanted to see if there was a difference if you gave people support while they were using the tools versus not, would you see some meaningful difference in their experience? And we found that it did. And I

Colleen Chien:

Would say that just to sort of build on what Miriam has said, what really surprised us I think is that the organic uptake was not necessarily equal in terms of what we know is that about 75% of legal aid professionals and lawyers are women. But in terms of who actually, when we took the survey among those folks who were in the legal aid profession, most of the use was by men. And so this sort of idea that build AI and it will come, everyone will use it and it’ll have this great equalizing effect. I mean, we can already see that there’s an uneven uptake of even familiarity or trust of the technology. So I think that was really interesting. But once we actually deployed the technology and gave folks access to the tools, there was really no difference in gender. People found it useful, they could do it, they could use it for a lot of different reasons, and that assistance actually improved the outcomes.

So again, I think this idea that there is a lot of promise, but we have to be thoughtful about how we deploy it, is something that came across that said when people did use the technology, they really appreciated it. And so that was very much in contrast to the headlines we’re seeing about Michael Cohen or the MATA case where there’s been a lot of caution in the legal community saying, this technology just isn’t really what fit for our purposes. We need to be really precise in how we communicate the law and how we engage in pleadings and we engage in legal research and that’s the case. But I think our pilot participants were able to show us how they were able to manage risks and sort of resolve this paradox of law needs high precision AI is not necessarily high precision, but yet people really want to use it. How did they do that? Well, they used it in ways where they can manage the risk. And I think that was important because I think a lot of legal aid attorneys don’t have time to try to figure out and double check all the work. But if it’s used in a way where precision is not that important, then it can be deployed in a way that actually saves people time rather than creates more work.

Cat Moon:

That is an excellent point, and that was a big point I made actually to the folks in Legal Aid who I was speaking with yesterday. And I can tell you that your research is already having an impact in the community because I’m fielding inquiries from folks in different legal aid organizations who are inspired by the evidence that having the support and training makes a difference. And so they are actively thinking about how can we intentionally engage with it by creating a platform and a foundation that gives folks the support and training that they need. And I think that is directly a result of your research.

Colleen Chien:

That’s so gratifying to hear. Thank you for sharing that work. We’re just really, really happy to

Cat Moon:

Hear so well done. And people are just coming to that themselves. Bréyonne, I want to ask you, you participated and so you were using the tool, you were one of perhaps contributed one of the case studies. So I am curious to know as a participant, what was a big takeaway from your perspective?

Bréyon Austin:

Definitely just going back to that support part, just having Miriam’s support. She gave us my team a training and I had no idea about the use of AI and the possibilities before this training. Learned a lot during that training, but then having Miriam participate in the clinic with us and show us how to use a lot of these tools was just really huge. So that guide and support was just crucial. I would say my biggest takeaway is again, in this new knowledge of how AI tools can be used to increase access to justice specifically for pro se litigants in a legal clinic setting because that’s my experience. And one part of that is just it showed me how attorneys could use these tools to increase both productivity and efficiency. So for example, my clinic, we have our pro bono volunteers and we have about an hour to meet with a client.

A lot of that time is filled up with the narrative, getting the narrative, why are you here today? And then that leaves a very small amount of time for issue spotting, clarifying questions, and then research and next steps, et cetera. So an example would be an attorney could spend 20 minutes or more before the use of these tools researching local rules and ordinances to help a client with a matter maybe drafting stuff for them. And using these AI tools, your results were almost instantaneous. Then we have some time to just, obviously I’m always going to say, please due diligence, check your results, check the work just to make sure we’re doing the best we can to support our clients. But during that, that then opens up the time for the attorney to talk and spend more time with the client and develop a connection. And that’s really important to us in our org. We’re very client centered, we are all human, and we want to connect in a human way. So having the use of these tools to speed up and make things more efficient gave us way more time to just have a deep connection with the clients. And then our clients are leaving feeling one that we’re able to help them in some way with their legal matter, but they’re also feeling heard, respected and connected.

Cat Moon:

That’s so interesting to me. So the technology is actually helping you better serve clients as humans. Absolutely. So you’re able to delegate some of the work that is not the immediate client facing part so that you can devote that really human touch.

Bréyon Austin:

Exactly.

Cat Moon:

And connection with clients. Yeah,

Colleen Chien:

I think Bréyonne’s comments also underscore the importance of the connection that Miriam made with Bréyonon in terms of being with her, showing her human to human, how these technologies can work, having the kind of confidence in them that they could be helpful, and having that trust that she already established from having a track record of being in community with and being partners with organizations like lawyers committee. So I kind of call it the Miriam effect in terms of pro bono work. And I do think it’s an interesting new model for how we do pro bono is if lawyers are being trained in their law firms and getting access to all these kind of great tools, they can work with the pro bono organizations that they’re working with and try to help them understand and be much more, just have access to the same tools.

Cat Moon:

Absolutely. And I’m inspired by the Miriam effect and actually am working on a pilot program along those lines, because there is just so much potential with this technology. I think that Bréyonon pointed out just an amazing example. And I’m curious, so is your organization going to continue to use and develop a process with these tools, give you these superpowers? I like to think of it.

Bréyon Austin:

Yeah, we’re still using it. We have our clinic weekly and we’re still using the AI tools. We check in with each client informed consent, very important obviously. So we check in with each client before using the tools. And I used them last week, this week, Monday. Yeah, definitely still. And we’re also looking into other ways to do it outside of the clinic setting as well.

Cat Moon:

So yeah, I love that you’re embracing it in that sense of experimentation and looking for other ways to leverage and use it. Well, I hope you will continue to find a platform to share what you all are learning through your organization, Miriam and Colleen, one of the important things I think you do in the paper that you wrote sharing this research is you identify opportunities and recommendations. So ideas for taking this and doing more with it, and one of those ideas you identify, I’d love to talk about all of them, but one of them is especially near and dear to my heart, and it’s something that has come up in many conversations here on Talk Justice, and that is the idea of communities of practice. And that would be a place in which Bréyonon and her organization and team could feed back into a group of people, other people similarly situated in different legal aid organizations and share learnings. And so I’m curious, I really welcome any of you to chime in on what that would look like and how we could make that happen. Because I think that has incredible potential and I don’t think it would be hard to make that happen. So I’d love to dig into that. Any thoughts on that?

Miriam Kim:

Yeah. Well, I think One Justice in California has been playing some role in that, right? They’ve played a real leadership role just in creating a community practice of pro bono work in general. And after we did the pilot, they invited me, I think through Bréyonne’s recommendation to speak to all the pro bono managers across the state of California about generative AI and the ethical issues. And then Bréyon and I and the founder of Gavel, Dorna Moine, who also supported our pilot, spoke at their annual conference to share our learnings. And so I think that is sort of the beginning steps of trying to use an organization like One Justice, which tries to cultivate a community in California among the legal aid organizations as well as pro bono managers at firms. So I could see the next step, and I don’t know, maybe this is already happening without my knowledge, but like a quarterly round table or a monthly conference call for those who are trying to advance the use of AI at the organizations.

I know they already have a number of monthly calls. So whether or not when Justice wants to take up that, cause I don’t know, but you could imagine another organization taking that up, whether it be the Justice Tech Association or a law school like Berkeley to take that up. So I think we really just, it may just need to be one person who’s willing to do that and sort of rally up the troops and see if there’s a critical mass of people who want to start. And if my dream would be that there would also be legal tech founders who are willing to support this effort and invite those who are part of this community of practice to pilot their tools for free and potentially for longer periods than they typically can maybe for six months or a year to really try to explore the potential here.

Cat Moon:

I think bringing legal tech founders into it is a fantastic idea. I agree. Colleen, would you like to add to that?

Colleen Chien:

Well, I was just going to say that I think that we underestimate how much demonstration effects and seeing other people, our peers do things that people we admire, do things really has a big impact. It’s not going to be because the open AI founders tell you to use the tool that you’ll use it, you’ll use it because your peer or your friend or somebody else said, oh my God, I just saved all this time doing this. And so I thought that it was really important when we asked folks, tell us your best use cases that they also told us how much time did they save and would they reuse it so that people could understand that some of these are better than others. We really want to be very sensitive to how people can have that first experience that is positive and then keep going with it. Because there’s a lot of experimentation. There is a learning curve to come up and how to use it appropriately. So I think knowing that there is the potential for closing up the gap to some extent that that’s at stake, I think we could drive more intentional efforts around this community of practice and make it more focused and be better resourced. So that’s our hope that that will happen.

Cat Moon:

Absolutely. Bréyonon, with respect to the community of practice, would you join another monthly call? It’s right, putting something, another thing on someone’s plate. I’m inspired by what you’re describing that your organization is engaging in, and so I just think it would be an awesome opportunity. So I’m inspired by the suggestions that you all have had so far. So I’m filing all that away. We might have to come back and have another conversation about this.

Colleen Chien:

One thing that we also try to do is really say that legal aid attorneys, they deserve the highest level of service and concierge level is what we wanted to provide. So there’s something about going on to another conference call or joining another training that’s not as responsive as I’m about to go meet a client, I don’t have time, help me, or I want to increase my impact 10 x, right? People that are in legal aid, they are super, they’re super people and they want to have big impact. So how can we help them on the spot in that moment? So I do think this concierge model where Miriam and a team of students, they were on call, they did office hours, they did, they responded to emails. It wasn’t just putting together the database, but it was really trying to think about how can we make you more productive and effective is appealing. So I don’t know, that’s not quite a community of practice, but something that’s really tailored towards, as you said, being client-centered and really thinking about legal aid lawyers as being the folks who really have the opportunity to, they’re the ones that can do the work, so we need to really lift them up and empower them. I think that is kind of how we were thinking about it ourselves.

Cat Moon:

Absolutely. I think it’s, and right, so how can you make that happen? How can you also connect people so that they can be learning from each other? I mean, I can envision your Airtable just growing into this BMI of searchable use cases that people can kind of slice and dice and figure out, here’s the challenge I’m having, and just see examples of how other people have tackled it. Yeah, so much opportunity there. And we also have to be intentional and cautious. Part of the important training is helping folks understand how this technology works so that they can use it appropriately and effectively. I always like to ask the question, what are the biggest concerns right now around using this technology for legal purposes? And that’s both the delivery of legal services and the operational stuff. Those things are perhaps lower stakes and not containing sensitive data and so forth. But there are reasons for us to be concerned. So what’s top of mind for you? All right Now in the concern category with the technology

Miriam Kim:

On my end, the biggest concern is hallucination, right? And you cannot guarantee that the information you’re getting is accurate. Now, that’s always true, even if you Google right, there are a lot of inaccurate things on the internet, but I think the output tends to come very quickly and come in a manner that exudes confidence, that gives you no reason to doubt its accuracy. And of course that poses risks if you’re a busy lawyer who doesn’t have time to verify, or if you’re a pro state litigant who doesn’t know how to verify. We certainly had use cases that we saw where there was an accurate output and the lawyer caught it, but I’m sure there are others where it’s not. So I think that’s why we tried to start again with our pilot with legal aid lawyers who do have the training to verify the accuracy, but also to explore use cases where legal citations aren’t necessary.

So in the legal clinic, for instance, that I was honored to participate in with Bréyonon, sometimes people came in needing a letter. It wasn’t a legal brief, it was really a letter that even a non-lawyer could write. So asking chat GPT to write that letter, whether it be to a government agency or to a landlord, or to just whoever it is on the other side of the dispute who probably wasn’t a lawyer either, that is something that chat G can do. So looking for those use cases where, yes, we know hallucination is an issue, but I’m not worried about hallucination if I’m just writing a letter to someone where I’m trying to resolve a dispute.

Colleen Chien:

I completely agree with Miriam. I think just to kind of answer the question a little bit differently, I think one of the concerns is around, again, thinking about augmenting legal lawyers, legal aid lawyers as one focus, but there also needs to continue to be attention on improving the quality of tools that are getting out there. So not only the quality of what legal aid lawyers are doing, but in my teaching AI class at Berkeley, and we were sort of doing some work and there are lawyers who are going online and just kind of reprinting Chachi BT answers, and there’s a small disclaimer saying you shouldn’t trust this. But there is a question of how these tools are being used and understanding also how we can regulate technology to make sure it’s of quality. So one of the recommendations we have in the paper is to really think about and realize that a lot of consumers are going to be using these technologies directly, not only just through and under the supervision of attorneys.

And so if we know that that’s the case, we need to think about how to certify or have some sort of quality standard that’s available for these offerings. And particularly for legal aid bots that are out there. And I know this kind of gets into the question of the unauthorized practice of law and the question of how much do we want to legitimize direct to consumer offerings, but I think they’re out there already. And so we need to be concerned about quality control in that context. And I will say from an access to justice perspective, even if we were able to double the productivity of every legal aid lawyer, we’d still have a huge gap. A lot of folks don’t know that IT help is available. They dunno how to get access to it. It’s not convenient for them. And so we do, I think, need to think about these what we call legally aid lawyer directed technology offerings.

So I’ll just kind of provide them my own research. I’ve done a lot of work with respect to expungement and what I call the second chance gap and these sort of paper prisons where one in three adults has a criminal record, many of them could get it cleared, but there has not been sort of easy ways to access expungement tools. So one of the tools that we highlight in the paper is Rasa, which was started by a legal aid lawyer, and she was able to really work with estate to get ADA and to understand who is eligible for relief and provide services. And in our interactions with her, she said that in the first year of their operations in Utah, which has kind of a looser regulatory environment with a sandbox, she was able to serve thousands, I think 10,000 people, whereas the entirety of legal aid in the previous year was in the hundreds. And so just understanding that there is this potential out there, we need to solve this unauthorized practice of law issues so that we can make and enable these technologies to be out there, I think is another thing that is important to pay attention to.

Cat Moon:

Absolutely. Well, I think one of the greatest opportunities is figuring out how we can scale this technology, to your point, Colleen, to be consumer facing, to really scale what right now is still a very one-to-one service model and is woefully inadequate and will remain. So there just aren’t enough lawyers in legal aid or otherwise to help all the people who need help. So Bréyon, I’m going to ask you actually as we wrap up a final question that combines, I think any concerns you have about the technology going forward as someone, again, boots on the ground using it with what you see as right now, the greatest opportunity before legal aid providers who are brave and courageous and wanting to experiment with generative ai.

Bréyon Austin:

Sure, definitely. And echoing all of the concerns about verifying the accuracy of results, I also think there’s a lot of fear about the use of new technology that’s definitely prevents people from exploring all of the possibilities. It could be firms, could be clients, could be anyone, and accessibility is also a major concern. Some of these tools just may be too expensive for legal service organizations and for self-represented parties. So those are some major concerns for me. And again, I’m always going to talk about informed consent and making sure everyone understands what’s going on and agrees to it. And then just moving forward, I think that it would be really great to see more pilot projects like the one that we participated in. I think that law firms have an incredible opportunity to partner with legal service organizations to further explore the use of these AI tools. At the end of the day, we all have the same goal of increasing access to justice. And so I think working together right now while it’s all being presented to us is extremely crucial.

Cat Moon:

Alright, so we’ll complete the final lightning round here. So Colleen, closing out, what do you see as the brightest and shiniest opportunity right now before us with respect to this technology going forward in the near term? I

Colleen Chien:

Think just getting it out and getting people to work with it and share their experiences as wide as possible at this point. And then we’ll kind of figure out the issues that it raises. But we just need to get the word out. We need to get the practice going.

Cat Moon:

Amen. What about you, Miriam? Any final words? Closing lightning round?

Miriam Kim:

Yeah, I just want to encourage people to share information. I think, Kat, you said it was interesting how we had our use case database. I have found that a lot of people do not like to share, especially if they’re at a law firm or otherwise, because either they don’t want their competitors to know what they’re doing. Or frankly, I think there’s some people who say they know how to use the tools, but they don’t really know. And so they won’t go into specifics about it. But our pilot participants were so amazing and so generous in their willingness to share with one another. And I really want to encourage the broader legal community to be thinking that way and to be putting on trainings to enable the next generation to also learn as they’re coming into the legal practice and just all work together to share our learnings to increase access to justice.

Cat Moon:

That is beautiful. Miriam, I think we should end on that note. That is a fantastic and inspiring sentiment. So thank you all so very much for having this conversation, for sharing your work. Please keep doing this important work. Please keep sharing about it, and I look forward to follow up conversations about your next project. So thank you so very much.

Colleen Chien:

Thank you.

Miriam Kim:

Thank you, Cat.

Bréyon Austin:

Thank you

Cat Moon:

Many, many thanks to my guest today, Colleen, Miriam, and Bréyon for engaging this conversation with us and sharing about this important research that they are doing and participating in. I think it is very informative and frankly inspirational. So I really recommend that you check out the paper generative AI and Legal Aid results from a field study and 100 use cases to bridge the access to justice gap. The paper itself links directly to a handy spreadsheet with a description of all the use cases, which I think apply across the legal services delivery spectrum and are not limited simply to legal aid. So I think it’s very informative and inspirational. So many thanks again to all of them for joining me. And thank you for listening to this episode. Talk Justice is brought to you by Legal Services Corporation and Legal Talk Network. If you like what you heard today, please be sure to rate and review the show and subscribe on your favorite podcast app.

Announcer:

Podcast Guest speakers views, thoughts and opinions are solely their own and do not necessarily represent the legal services corporation’s views, thoughts, or opinions. The information and guidance discussed in this podcast are provided for informational purposes only, and should not be construed as legal advice. You should not make decision based on this podcast content without seeking legal or other professional advice.

Continue Reading