[Music] welcome back the Department of Veterans Affairs is showing its healthc care providers all over the country the potential of artificial intelligence it's doing it though differently than you might expect Jonathan album is federal CTO and principal digital strategist at service now he's former Chief Information officer at the Department of Agriculture Jonathan welcome it's great to see you again VA is doing things similar to what you and I have talked about in the past they're calling their concept trustworthy AI the way that they're trying to demonstrate to both their care providers and their end users their their uh their clients what the AI will do for them the second thing they're doing is they're going to those care providers and they're saying what do you do a lot of that you would like to do less of and what are you not doing enough of to provide service uh to your patients is that what you're seeing other agencies doing that definit cont tracks with what I'm what I'm seeing and hearing you know these um these generative AI technologies that are built on large language models have some uh risk around data privacy and and reliability when you when you look at these commercially available Technologies chat gbt open AI these these things have tremendous value but if you play with them and you can see the results that come back uh you might question them and you know if you can't trust the result how can you apply that in a government setting when you're trying to serve uh serve the public so you know what I think will occur over time is that these these technologies will make it into government but they're going to be based on smaller large language models and these these smaller large language models you know will be built on maybe industry specific data or agency specific data or data from a technology provider that has lots of information about how technology is used and what makes it effective and we can use AI tools on on top of those data sets and those results will likely be a lot more trustworthy and reliable and if you know organizations create these large language models responsibly then they can be audited and people can have the transparency that makes this technology something that uh can be trusted and and and used in a in a setting again where you're you're providing a critical service to to a member of the public will agencies need to have different specialized tools to go with specialized data or will they be able to use the tools that people are using more broadly um ideally they're going to be able to use the tools that they have and have those uh those data sets feed into them I mean when you look at technology platforms that are out there many of these technology platforms have had AI built into it for a long time and whether it's around predictive intelligence or some kind of document intelligence these tools have been leveraging AI capabilities in federal space today I think what's different is um some again something like chat gbt is is sort of like Netscape in the early days of the internet the internet was around but Netscape made it accessible chat GPT makes generative AI accessible so people now can start thinking about how to apply it so the trick is to be able to use it within your existing technology construct so the Val so you get more value from these Investments you've already made and I I I I have a high level of confidence that's the way that these tools um provide real value in agencies and you know make it into day-to-day use cases I like the Netscape analogy for the way that the the potential exists to harness the technology and make it accessible to people on an everyday basis what lessons can we learn from that technology progression the browser explosion allowed people to access the internet compared to what we're seeing because it's it I think you're exactly right that that's what happened with generative a well I I think it goes u i I'd say there's there's something in what you you led with earlier about the VA asking users what they need as opposed to thinking okay now we have this technology I'm just going to deploy it and it's going to solve our problems you know it's always more complicated than you have technology well what are you trying to accomplish so the fact that VA is asking those questions I think is um a healthy sign that they can be successful adopting these Technologies it goes back to you know what is the data that you have in your organization how does that data flow through the organization what are the processes you you rely on and ideally agencies have spent a lot of time digitizing processes today and you know we add AI technology on top of that and you take the data that you know you take these processes these workflows that are uh that are that are digitized and providing real value and then you you know you elevate them using gener of AI technology to make them more effective and make people more productive less than a minute left what are the next steps after finding out what employees tell you they do too much of and don't do enough of what do you do after that Jonathan well I think we have to have conversations with both employees we have to have conversations with our customers and their agencies to explain that these Technologies are going to be used because there's a lot of fear out there there's a a lot of questions people are are skeptical about you know um Ai and its long-term utility so if we can have the right kind of dialogue and we can get people on the same page about the things AI will do to provide value to make somebody more effective to make a citizen get the service they want faster a a federal employee be able to serve that person faster or the agency to operate more effectively and the Technologies are providing value to taxpayers I think we're set up to be successful long term as a as a the federal it Community Jonathan album great to see you as always my friend thank you thanks Francis you can learn more about