In this 5th episode of the special AI-focused mini-series, guest hosts Pauline James and David Creelman talk with entrepreneurs in the HR Tech marketplace. This discussion is focussed on practical use cases for AI and offers up insights from four experts – Boyd Reid, Fuad Miah, Mike Hall, Ania Garczyńska and Shingai Manjengwa – to discuss how AI is driving innovation in HR products and departments.
Tune in and Discover:
We do our best to ensure editorial objectivity. The views and ideas shared by our guests and sponsors are entirely independent of The HR Gazette, HRchat Podcast and Iceni Media Inc.
Feature Your Brand on the HRchat Podcast
The HRchat show has had 100,000s of downloads and is frequently listed as one of the most popular global podcasts for HR pros, Talent execs and leaders. It is ranked in the top ten in the world based on traffic, social media followers, domain authority & freshness. The podcast is also ranked as the Best Canadian HR Podcast by FeedSpot and one of the top 10% most popular shows by Listen Score.
Want to share the story of how your business is helping to shape the world of work? We offer sponsored episodes, audio adverts, email campaigns, and a host of other options. Check out packages here.
In this 5th episode of the special AI-focused mini-series, guest hosts Pauline James and David Creelman talk with entrepreneurs in the HR Tech marketplace. This discussion is focussed on practical use cases for AI and offers up insights from four experts – Boyd Reid, Fuad Miah, Mike Hall, Ania Garczyńska and Shingai Manjengwa – to discuss how AI is driving innovation in HR products and departments.
Tune in and Discover:
We do our best to ensure editorial objectivity. The views and ideas shared by our guests and sponsors are entirely independent of The HR Gazette, HRchat Podcast and Iceni Media Inc.
Feature Your Brand on the HRchat Podcast
The HRchat show has had 100,000s of downloads and is frequently listed as one of the most popular global podcasts for HR pros, Talent execs and leaders. It is ranked in the top ten in the world based on traffic, social media followers, domain authority & freshness. The podcast is also ranked as the Best Canadian HR Podcast by FeedSpot and one of the top 10% most popular shows by Listen Score.
Want to share the story of how your business is helping to shape the world of work? We offer sponsored episodes, audio adverts, email campaigns, and a host of other options. Check out packages here.
Welcome to the HR Chat Show, one of the world's most downloaded and shared podcasts designed for HR pros, talent execs, tech enthusiasts and business leaders. For hundreds more episodes and what's new in the world of work, subscribe to the show, follow us on social media and visit HRGazettecom.
Speaker 2:Hello and welcome to the HR Chat Podcast. I'm Pauline James, founder and CEO of AnchorHR, and it's my pleasure to be your pod host today. Along with David Krillman, CEO of Krillman Research, we're partnering with the HR Chat Podcast on a series to help HR professionals and leaders navigate AI's impact on organizations, jobs and people. Many of the people we talk to are from big companies or technical experts. We're going to start today with the change of pace by speaking to a Canadian entrepreneur who hopes AI can improve our ability to help employees get to and from work. Boyd, could you take a few minutes to introduce yourself?
Speaker 3:Yes, my name is Boyd Reed. I am the co-founder of Hopin Technologies. We are a logistics software platform that connects employers, retramptation operators and efforts to help employees get to and from work with use of shuttle services. So we help employers retain more talent by making it easier for their employees to get to and from work, but also recruit more by giving their job site more accessibility and transportation options to get to and from work as well.
Speaker 4:Boyd, what is the technical challenge in organizing transportation that got you interested in AI?
Speaker 3:That's very complicated. When employees are located kind of north, south, east and northeast everywhere, it's hard to coordinate routes, and then it's also even harder to coordinate them with the transportation options that are potentially available to you. So what we've been able to develop at this particular stage is algorithms that look at logistics in terms of what we call relational data, so looking at each employee in relation to their colleagues, in relation to their shift times that they're working, in relation to the transportation options that are currently available. We have an algorithm that looks at, looking at all these particular points, identifying gaps and then spitting out solutions to how we can address this particular problem there. We don't have a fully functioning machine learning tech, but we're in the process right now of developing our algorithms and actually working with data scientists and professors at Canadian universities help us develop out a machine learning technology over time.
Speaker 3:The biggest challenge right now is that people's schedules tend to be all over the place and as we continue to build up and build up and servicing more and more people, we're not going to be only servicing one or two companies at a time.
Speaker 3:We could be dozens or multiple companies, our entire cities, our entire areas. What we're not able to do right now, because the technology that we have isn't there yet is coordinating 50 companies all at once utilizing the same transportation resources, looking at all the varying amount of times, the very amount of shift times when people need to get into work. That point there is where there's too much for the human brain to be able to coordinate on those things, unless we want to just sit there at the computer all day. So what? The AI and these algorithms that we're building towards? That machine learning technology will allow to be able to process all that information a lot quicker by just taking even just excel sheets from shift times where people are located and all that information to split it in there, and then it will be able to process it at a higher rate.
Speaker 2:I can see how the AI scheduling of shuttles can make these efforts to support organizations and employees overcome challenges with commutes to work. What is your long term vision?
Speaker 3:One of our bigger visions for what we're able to do. One of our corny sayings that hop in is that transportation is the vehicle to opportunity. What we mean by that is that transportation literally has access to everything healthcare, jobs, play, family. So by being able to utilize this type of technology, we'll be, in essence, be able to support public transit infrastructure by utilizing private transportation infrastructure and using our AI to logistically coordinate all those moving pieces.
Speaker 3:Currently, we're not at the capacity for that but that's where we're in essence building towards with hop in, but we started with the business because you know that has huge economic impact on our communities. There's also consistency in that and then from their building, from the businesses being able to support the community at large.
Speaker 2:Boyd, I'm curious, as an entrepreneur, as a business owner, as someone who's in the HR vendor space, who's embarking on this journey of working with the sophisticated systems that you've built and upgrading them with these advances in technology, what advice you have for those of us on the other side, whether we're looking at how we can innovate within our own operations, whether we're partnering with vendors.
Speaker 3:Yeah, what I like to say about that is I'm a non technical founder, which means that I don't code. I don't have any programming background. Everything that I'm at right now is all self taught. Now that I still can't code, I can't build apps and utilize these things. I utilize our resources and experts in that particular space, but my advice, because of my particular background, is that it takes steps.
Speaker 3:So, once you have a particular vision, it's not about jumping to how AI or tech can solve it right away. It's like how can you solve it even without tech? The tech is in itself as a tool to be able to accelerate and scale, so it shouldn't be the end product. It should be a way of accelerating and are being able to help scale a particular solution.
Speaker 3:I think that mindset is really really key, because people like to jump towards tech as the solve all right away, but really all it does is solve a problem that has been around forever, or making it more efficient, or anything like that. I would look at seeing how can I solve the problem now on a small scale and then, once you're able to solve that, prove a concept and build out then from there, how can I utilize technology to be able to accelerate that or scale it out. Some people think about it in reverse, and that's where really the downfall is and disconnect. But when you have a strong mission, like what we have up in technologies, which is, how do we get people to work, we can do it with tech or without tech, but the tech allows us to do it better, at a more efficient way and at a scalable rate as well, and that's the mindset that I would very much encourage and advise around people is just think about the solution for the implementation of tech.
Speaker 2:Boyd, as we consider the significant advancements in technology and AI and automation as of late, it comes with a lot of upside, which you've described, what you're exploring and developing within Hoppin. I'm wondering if you can share what you also see as risks that need to be considered and managed.
Speaker 3:I want to see us on like in the next galaxies and traveling space and flying cars and all those things in my lifetime selfishly, but at the same time I understand that what needs to be done in terms of pressures, like rigorous testing to show we're putting in stop gaps and things in place of that. We know that people can utilize this way, but we have stop gaps to stop people from utilizing it that way. And I think right now, because tech advancement is advancing way faster than any policy or regulation can keep up with, they're able to play in this gray space where they can just say oh, and kind of throw their hands up in there and say I didn't think that it would happen that way. But I think generally as a community around the tech space, we have to be mindful of and take that moral eye ground of putting in rigorous testing to help stop any particular risks or at least mitigate any risks that potentially happen from all this technological advancement.
Speaker 2:With these podcasts, we are often focusing on AI technology, but perhaps we can wrap up by talking about the human side.
Speaker 3:One of my favorite stories to share on that is with one of our clients, Maple Osh Farms.
Speaker 3:This employee has the ability to drive into work, but one of the reasons why they use our shuttle is not out of not being able to get to work, but they work to late shift and they have two young kids and one of the things that would happen is that they would get at home and pretty much fall asleep immediately because they'd be exhausted and they miss out that morning time of getting them ready and all those memories that you make on that.
Speaker 3:So what they were able to do on our shuttle was actually take a nap. Take a nap so that they're alive and ready to deal with their kids running around in the morning. And it was just something that was so small that you don't even think about it because it's just something that it's just a very particular, unique case. But it shows not only saving people time in terms of the community, but being able to provide them an option for them to relax a little bit and get into work, what that can do in terms of impact while in their life. And that's one of my favorite stories to share because it's something that you never think about, but it just makes a lot of sense. Being able to nap so you can be awake for your kids in the morning is something that is just so powerful in that narrative.
Speaker 1:Ascendium is on a mission to make education and training beyond high school a reality for learners from low income backgrounds, and we're pleased to support this episode of the HR chat show. Earning a credential certificate or degree has the power to transform a person's life. It can lead to financial stability, a fulfilling career and a stronger community. Ascendium works to create opportunities for learners from low income backgrounds to achieve that transformation. Learn more at ascendiumeducationorg. And now back to the show.
Speaker 2:We heard from a business looking to use AI to partner with organizations to solve challenges related to commuting. Now let's turn to a technical expert who helps businesses build AI tools Fuad Mia. Fuad is co-founder of Unkennie Lab Limited and is deep expertise in the application of large language models.
Speaker 4:The scenario I'd like to explore is that imagine you're an HR leader and you're somewhat familiar with generative AI and large language models and your team has been playing with chat, gpt and so on, and you've had various vendors come along and offer you solutions. But then some of your team is saying you know, we really need to build our own application using generative AI, for example they might have an idea about. We want to build our own custom chatbot for HR. And then the question becomes just in general, when I get these proposals, the first question is how do I decide if this is a practical idea at all? How do I decide if I should even consider this potential application?
Speaker 5:That's a great question. I think one of the very first things businesses have to understand is what is my golden use case, even before going to that staff, getting a proposal or getting into meetings with vendors, and I think that's where the biggest challenge is right now. Understanding the value of AI is quite easy, but how you leverage it without breaking your bank, selling a kidney that really overnight will transform your organization. And I think that particular phase needs to be discussed, ideated, brainstormed, and when you're ready to take the next step in terms of validating that idea, that's when you have to start conversation with these vendors. But in between that or before that, you may have to talk to experts, subject matter experts to validate your idea, before you even think of a POC with vendors of these large language models.
Speaker 4:And you've mentioned that, after you've validated the idea, you need to do a POC, which is we all love acronyms proof of concept. What's involved in doing the proof of concept before you make that big investment to actually build it?
Speaker 5:Now that's an amazing question, david, and the good news is not some, I would say most of these vendors, including Google, microsoft they have built interfaces where you can test your ideas. You can actually enter a fraction, let's say, of your data set and then, using different, let's say fine-tuned models for different objectives, like question answer versus summarization versus insights from a document you could take this out of the box, fine-tuned models for different objectives, and see how they perform against your attraction of your data set before you build an automation. And I think that should be the first step. When you have a validated idea and then you try to test with your data set, what are my outcomes like? What should it be like?
Speaker 5:There are other nuances we have to think about whether your idea or use case requires fine-tuning of a model that exists, or it requires just few short learning, or it could be just even out of the box, for example, tasks like summarization and so forth. Fine-tuning, when you come up with a use case that requires substantial customization for a specific task in a domain like HR, let's say, that would require a little bit more than few short learning. Few short learning just let me define it a little bit where you show, just like the name suggests. You show a few suggestions and answers or a few examples of your outcomes and then it predicts from those few examples versus fine-tuning, where you would be using a larger internal data set and its outcomes and so forth. And that requires data cleaning. That requires understanding whether my data set is well diversified I'm not showing all possible examples or not, whether I've taken some of the relevant nuances out that will impact negatively.
Speaker 4:Now, obviously you're going to be as the HR leader. You're going to be relying on the big vendors who provide the base models and tools, and you're probably going to be relying on other vendors or consultants to help you through the project. What kind of talent do you need internally? Because all this is quite new and things like few short learning. The HR leader might not. Even though you've explained it, it might still be a little bit vague to them. So what kind of talent would you recommend people hire or develop internally?
Speaker 5:Once HR leaders understand the use case, the talent you would be needing is not necessarily someone who is very technically well versed in AI itself.
Speaker 5:If you need someone who understands your use case and understands the overall spectrum of AI, that would allow him to hire the right talent for the right models that he needs.
Speaker 5:Ai is a huge spectrum and in this case, if I could, I would say the skill sets would be required someone who understands data science, someone who understands the very high level of different use cases and in different models. So, for example, what is classification, what is sentiment analysis, what are the artifacts that you can extract from different interactions and data and how you can take that into business context. So I would say it is quite a unique skill set that you'd need to understand the high level of your business HR domain expertise studies and who also has an overall understanding of various models out there that's available and different skill sets that's required to work with those models, and as well as understanding of integrations and that idea is quite mature now the interoperability and integrations. So you need one person who understand three factors AI, data science, integrations and domain expertise knowledge as well not as in depth, I would say but who has an understanding of the use cases in this particular business?
Speaker 4:And I want to underline what I took away from that, which I think is very important, is that my job as chief HR officer is not necessarily to hire the people who write technical skill sets. It's to hire the person who can hire those people who write technical skill sets. So I want to hire that person in the middle who knows a lot more about AI than I do, still understands the business, but I'm not going to expect them to know all the technical details of these different application. They will know enough to hire the people they need to drive this forward. That's great advice.
Speaker 1:And now back to the show.
Speaker 2:If you are ambitious, you may, like Boyd, develop your own AI tool. Perhaps you'll work with someone like Fawad to make it happen. But even if you don't want to take on that bigger project, they're off the shelf AI tools that can help. One thing that exhausts HR pros is screening resumes, and frankly, we don't always do it well. Let's talk to someone about what AI can do with this one task.
Speaker 4:Mike Hall is co-founder of Lighthouse Hiring Innovations and, as you might guess from the company name, mike is working to apply technological innovations to improve hiring. Thanks for joining us, mike.
Speaker 6:Hi, I'm Mike Hall, CEO and co-founder of Lighthouse Hiring Innovations. We're using AI to make the hiring process, and specifically that candidate review process, more fair, objective and honestly human. As an example, if we think about, let's take a sales role and account executive role and if we post that out, we're likely to get about 250 applicants on average. Now, someone's going to have to review that, and whenever we have humans reviewing resumes, there's a few things that crop up, especially at that volume. One. There's bias, and unfortunately we all try to avoid this, but it's really hard to, and so we have unconscious bias and relevancy bias that's introduced at resume reviews. You have a massive time constraint when it comes to reviewing all of those resumes, so lots of time spent to be able to really do a good job of those reviews. And then three, you also have review fatigue. So your first review is going to be a lot better than your 60th.
Speaker 6:And if we think about the tools that exist today you know ATS systems. They do definitely help. You have keyword searches that are out there, but we all know that context within the resume is really what helps us understand a candidate, and keywords just don't usually help with that. So where we come in as we leverage our AI models and our natural language processing to really understand a candidate just like a human would, and really factor for all those inputs that the hiring manager gives us and what they're looking for as an example for that account executive role. You may be looking for prospecting as a core competency and the candidate may have put in the resume context which really truly means prospecting, and a human could identify that, but they never actually said the word. What's cool is that our models and our natural language processing can actually pick up on that and provide that label of prospecting to that context, knowing that that's something that the hiring manager is looking for and so what this means.
Speaker 6:If we think back to that problems with the human review side of things. Now, one, we're not introducing any bias because ultimately the models don't look at anything that is a biasing variable, like name or email or address. Two, we're saving a lot of time because all of these reviews are now totally automated. And then, three, there's no review fatigue. There's a ton of consistency across all of these because computers don't get tired, and ultimately this really means is that now you have a much more objective and fair based review process, because you do have that consistency and you're not introducing those potential biases. Really, that's how we're making it more fair, objective and easy with AI.
Speaker 2:Next up in our discussion. I'm happy to connect with Anya Garzynska, who co leads a hot European HR startup based in Poland. Her company has taken analytics of collaboration to a new level by adding in capabilities based on chat, GBT. Anya, can you give a brief introduction to who you are and what you are working on?
Speaker 7:My name is Anya Garcheńska and I'm a co-founder of Network Perspective. We've built an analytical tool that helps to improve efficiency by measuring and changing collaboration habits. This is the app which shows teams how to remove constraints to growth by paying back organizational debt at the same time as the team keeps doing the day-to-day work. So the app serves as the first step to start a chain reaction that results in practical, measurable steps towards becoming smarter and even more adaptive. Why is it so important? Because up to 85% of our work time is consumed by meetings, chats, emails and context switching, which, according to any research, drain productivity and burn people out. According to Gallup, $322 billion yearly is the cost of turnover and lost productivity globally due to employee burn-up. The issue is quite big.
Speaker 2:Where do you get the data to do your analysis?
Speaker 7:So how does the app work? We use the data that already exists in company's collaboration tools, such as meetings and emails in Google or chat since lag, and based on that, we provide teams with insights related to the ways of work and we suggest smart actions that can be taken to improve. The important thing is that we process the data with strong ethics up front. There are high security levels, data is hashed in the flow, there are only group reports and there's no content download. The first building block of the app is the data itself, which allows leaders to sense the communication and collaboration dynamics both within and across teams. The second building block is intelligence. We automatically provide teams with insights describing ways of work in terms of meetings, deep work and context switching habits, but also in terms of collaboration patterns within and across teams. The third and the last building block of the app are insights and actions, which help teams to change who is talking to whom, about what and when.
Speaker 4:Why did you choose to focus on collaboration?
Speaker 7:Transforming the flow of information within a group is the fundamental dynamic on which all other changes depend or are driven by. So the result is that the app is reducing time spent on meetings and context switching to let people save minimum 40 hours monthly per team. That's an extra week of work, right.
Speaker 4:Can you dive into the AI component of this tool?
Speaker 7:How do we apply AI models into the tool? So we use chat GPT to summarize the reports and present key findings based on the team's data and to assist teams towards working smart. The assistant knows your team as fully reported, with all team metrics, is included in the prompt, and, at the same time, the assistant knows best practices shared by the tech community and top companies such as GitLab, buffer, spotify or Dropbox. Chat GPT can link you to them and answer any questions related to your team's data, insights and best practices already used by different tech companies. We've also built a neural network that, based on the data, describes events in a week of work, for example, meetings, chats and emails, but also based on interactions the specific personas or teams is sensing signals, helpful in predicting burnout or isolation among teams. As the network perspective app provides teams with signals sensing how the work is done and guides teams to are working smart, it begins the process of intelligent changes and adaptive reactions to one becoming smarter and more adaptive.
Speaker 2:It is interesting how AI can take us that last step, from analyzing data to providing insights. Thank you for sharing your work with us. After hearing from entrepreneurs on how they are advancing their offer by leveraging the power of AI, and from what on the technicalities of AI and HR, it's clear that AI has diverse applications, but with these technological advances comes complexities. Shingai Mnjangwa, who we met in an earlier episode to discuss how organizations can harness AI effectively, rejoins us today to discuss how AI can be utilized in HR for efficiency and improvement, while maintaining our ethical standards. Shingai is the head of AI education at ChainML, serves as a board member for the Institute on Governance and is the founder of Fireside Analytics, an organization dedicated to providing education aimed at fostering data and tech literacy. Shingai, can you share your thoughts on AI and HR?
Speaker 8:So how can HR people use AI in their functions to advance what they're doing? Well, it starts with being able to define your objective functions. Hr people need to be able to say I'm trying to maximize something, I'm trying to minimize something or I'm trying to optimize something. So a maximization goal would be I'm trying to maximize the employment survey score for a particular objective, or I'm trying to maximize the retention for a particular band of roles. Those are the types of objectives that HR people would come up with and from the objective we would then go and see what machine learning tool, algorithm or AI we can use to be able to do that. It could be I'm trying to reduce the amount of time it takes me to go through many resumes in my recruitment process. That's an objective function that you can set up to say, reduce the amount of time, and so we would then look at what automation or what tools can you leverage to help you do that? And if natural language processing can read resumes for you, then you can use that certainly to do your first layer of screening for different roles. So in that sense, hr people can use AI for many different things. But I couldn't say that to you without highlighting the possibility of bias.
Speaker 8:A good friend of mine, paranas Chavani at Georgian Partners, did a fantastic case study where turn it in, a popular tool for screening cheating at universities, was disproportionately flagging non-native English speakers for cheating. So those are false positives. It's saying that someone is cheating when they're not and it was flagging non-native English speakers. Now why the algorithm is actually doing that, we can hazard a guess. We can say maybe those students were using translation tools, maybe those tools just their sentence structure is going to be different because they're translating from a different language as they type.
Speaker 8:Whatever the good reasons are, that's a problem Because you're getting false positives in a cheating scenario. But imagine in a job application scenario. There are even some roles that, as Shingaimanjengwa, I would never even see, based on the tools that we're using to be able to promote and highlight and advertise those jobs and roles. So, in as much as we can use these tools to make our jobs easier, I have to say, right next to that, the caveat that says we have to be testing for bias, we have to be conscious of bias and we have to be promoting diversity in our staff, in our communities.
Speaker 8:I know that's a self-serving thing for me to say as the Black woman from Africa, but it is true that diversity does empower our teams and it's, at the very least, I also say it's a risk mitigation strategy. In a world that's so dynamic and changing, having these different perspectives on our teams is a superpower at the end of the day. So, yes, hr, people can leverage AI, but please be mindful of bias, because, at the end of the day, we rely on you to protect us as a workforce.
Speaker 2:Shingaimanjengwa. Building on that, wondering if you also foresee ways in which AI can support us in overcoming bias, which is so inherently human.
Speaker 8:So we have bias as a result of technology, as a result of models. However, models can also help us to detect and mitigate bias. So in your deployment of a classic machine learning model, whatever it's doing, you get a certain amount of error. Typically, they don't predict 100%, and if they do, you should be very skeptical about that model. So when you're deploying your machine learning models, you typically have some form of error, and so I always encourage data scientists to please go and investigate your errors. You have to see, is there anyone who is disproportionately any clusters or groups disproportionately coming through in your errors as you deploy different prediction algorithms or different machine learning models, whether it's in an HR context or any other broader context that involves people, and in addition to that we have mathematical approaches to reducing bias. So there's we can say something like balancing your classes, making sure that you have in your training data sets the ones that determine your prediction and how well it works In your training data sets, making sure that you have different types of people in there that are representative of the distribution that these models are going to be deployed in. So just speaking candidly, if we go and retrain a data set in Kenya, say for a computer vision and people are starting to use tools that do this. You want to do a test, some sort of an onboarding test, or your technical staff? Sure, let's use a computer vision algorithm. If we go and train that in Kenya and then we try and deploy it in certain parts of the United States, you might end up having a tool that doesn't quite land because the training set that was used to develop that tool doesn't reflect the community that that tool is going to be used for. The best way to put it is that the community that you're building for has to be incorporated in the model itself so that you end up having a more fair model. But in addition to that, even if you balance your classes, bias can still come up. That's why I was advocate for testing, testing, testing. You have to actively test for bias. Anytime that you're deploying these models and you're dealing with people Any sector, whether it's human resources, healthcare care, work in general where we're starting to use artificial intelligence or more automated tools, we have to take extra steps to ensure that we are being fair and that no one is being negatively impacted by the tools that we're planning to use.
Speaker 8:One point I'll make is that we all come at artificial intelligence and technology from different backgrounds and always joke that on university campuses we always put social scientists on one side of the campus and engineering and science and math on the other side of the campus to make sure that there's no contamination of ideas. And we've all brought that into the workforce and we've brought that into our jobs. So sometimes thinking that, oh, now I need to start learning AI, now I need to start learning about machine learning and algorithms. I didn't set out to do that. That's not my strength. Some of those ideas are limiting and I think the pervasiveness of the technology is such that we don't get to take that position anymore.
Speaker 8:Everybody has to learn. If HR was your undergrad degree, working in the field for many years, and you just don't feel like you want to take up this mantle of technology, I would say that you need to rethink that. Never stop learning. We all have to evolve and keep upskilling ourselves and keep ourselves relevant. Let's just throw down some of those shackles and be open to learning, and it's one step at a time, right? That's what education is Meeting people where they're at. So start with what we know and then take small incremental steps to where we need to be. That's a message for HR people, but I also give that to whether I'm teaching students and high school level or college level is we all come at it from different places. Let's just be open-minded. Never stop learning.
Speaker 2:Tremendous advice. Thank you, Shanghai.
Speaker 4:In wrapping up today's episode, we thank our guests for their insightful contributions to our understanding of AI and HR. Their expertise reminds us of the balance needed between embracing AI's efficiency and being mindful of the ways we can identify and mitigate risks. Join us next time as we continue to explore the intersection of technology and HR. Until then, keep learning and have a great day.