Learning Library

← Back to Library

Hide-and-Seek: Uncovering AI Assets

Full Transcript

# Hide-and-Seek: Uncovering AI Assets **Source:** [https://www.youtube.com/watch?v=vI_2LSc6NCg](https://www.youtube.com/watch?v=vI_2LSc6NCg) **Duration:** 00:08:51 ## Sections - [00:00:00](https://www.youtube.com/watch?v=vI_2LSc6NCg&t=0s) **Finding AI Data in IT Silos** - Hillary Hunter likens locating data and models in tangled IT and cloud environments to a game of hide‑and‑seek, emphasizing that unified, consistently designed infrastructure is essential to eliminate silos and accelerate AI deployment. ## Full Transcript
0:02I think all of you will know the Beloved 0:04childhood game called hide-and seek I'm 0:06sure it goes by different names around 0:08different parts of the world but in 0:09today's era of generative AI your data 0:13and your models are going to be two of 0:14your most important assets however 0:17finding those assets in a complex it 0:20environment sometimes can feel a lot 0:21like playing a game of hide-and seek and 0:24it really doesn't need to be that way 0:26you don't want to spend months trying to 0:27make sense of your different it and 0:29Cloud environments and wrestle through 0:31different data governance systems before 0:33you can actually access and gather the 0:35data that you need to become productive 0:37with AI because there's no AI without 0:40data after all but a disjointed chaotic 0:42it environment creates silos that make 0:44it difficult to seek out and collect the 0:47data that you need to train tune and 0:49leverage your AI welcome to AI Academy 0:52my name is Hillary Hunter and I'm the 0:53chief technology officer of the IBM 0:56infrastructure unit I oversee the 0:58technical strategy and the development 1:00and deployment of platform capabilities 1:02for our clients in our hardware systems 1:04and Cloud families and these days my 1:06mission is to identify how we can use 1:08infrastructure and platform capabilities 1:10to enable Ai and maximize its value for 1:13our clients that work starts with making 1:15intentional consistent architectural 1:18decisions that eliminate silos and 1:20strategically bring data and AI together 1:23each component of your it infrastructure 1:25should fit together like a puzzle but 1:27gaining access to your data or finding 1:29your models shouldn't feel like solving 1:31one a decade ago we had the first round 1:34of ai's emergence in the broader 1:36commercial space it was an ERA marked by 1:38extensive experimentation lots of proofs 1:41of concept but relatively limited 1:43adoption early AI couldn't do more than 1:46identify objects in a picture and add a 1:48caption perhaps or deliver responses to 1:51predetermined questions built with 1:53rule-based systems these use cases 1:55weren't pervasive enough to justify 1:57adoption for many Enterprise 1:59organizations s those who did adopt 2:01could find the latest AI capabilities 2:03usually in the cloud and only in the 2:05cloud in order to use it adopters had to 2:08ship all their relevant data to the 2:10specific Cloud environment that housed 2:12that Ai and with so few architectural 2:14options available organizations had to 2:16put a lot of trust in their Cloud 2:18providers and in many cases invest in 2:20the high cost of moving large amounts of 2:23data to the cloud Ai and Cloud 2:25architecture look very different today 2:27gen AI is powered by a class of deep 2:30learning algorithms including Foundation 2:31models large language models which 2:34enable rapid solutions to be built in 2:36the areas of things like natural 2:38processing simplifying Enterprise 2:40operations and enabling Enterprises to 2:43take application modernization and 2:44workflow automation to the next level 2:47gen can perform important tasks on the 2:50Enterprise level like creating and 2:52understanding code authoring marketing 2:54materials detecting fraudulent activity 2:56and extracting insights from 2:58unstructured Text data 3:00with commercial use cases multiplying by 3:02the day and the growing potential to 3:05maximize business value organizations 3:07are eager to start adopting and 3:09deploying AI rightfully so however Cloud 3:12technology has also expanded 3:14significantly we're now in the era of 3:16what we refer to as hybrid cloud in 3:18which Cloud Technologies are available 3:20across the multicloud landscape and 3:22spanning into un promises environment so 3:25traditional it Estates private Cloud 3:27public cloud and even the edge become a 3:30part of the hybrid Cloud conversation 3:32critical business data often resides on 3:34premises in the data center it's siloed 3:37but it's kept secure which is why many 3:39organizations store their most sensitive 3:41data on servers on premises the problem 3:44is that traditional management of 3:46servers without a unifying platform 3:48approach lacks the flexibility needed to 3:50integrate and create AI at scale a 2023 3:54study IBM conducted with the Harris 3:56organization recently found that 65% of 3:59organizations are using a combination of 4:02on-premises Hardware with multiple 4:04private and public Cloud capabilities 4:07creating a hybrid environment and 4:09complicating their it landscape many 4:11organizations have fallen victim to what 4:13we can call a Franken Cloud it's a 4:15monster of each organization's own 4:18creation the Franken Cloud essentially 4:20is a concept that says you have a very 4:22muddled incidental hybrid Cloud 4:24architecture it has all the right 4:26components to meet the definition of 4:28hybrid it has some cloud CL on premises 4:30and public Cloud usage but it lacks the 4:33structure and intentionality to be a 4:35hybrid Cloud a consistent environment 4:38for example a large Enterprise may have 4:40several business units working with the 4:42same data with no centralized place to 4:45access and analyze this data each 4:47business unit creates its own cloud 4:49environment and moves a copy of the data 4:52to that environment the organization is 4:55then left with umen isolated Cloud 4:57deployments that increase data silos 5:00increase storage costs increase security 5:03risk and effectively add architectural 5:06complexity to the overall landscape 5:08trying to run and build AI models in so 5:11many different locations only worsens 5:13these problems so what's the solution 5:17well it's not to choose between on 5:19premises infrastructure where you may 5:21have limited access to AI capabilities 5:24or choose between Cloud environments 5:26that put your data privacy and Security 5:28in a totally different comp context 5:30instead businesses need to adopt an 5:33infrastructure strategy that's hybrid by 5:35Design and work with an AI vendor that 5:38offers choice in AI deployment location 5:41a hybrid by Design approach eliminates 5:44the challenges born from random acts of 5:46cloud usage by establishing consistency 5:49across the various functions tools data 5:51sources platforms and systems that make 5:54up your whole it environment you're not 5:57simply adopting hybrid technology you're 6:00making intentional architectural 6:01decisions that unify every part of your 6:04business you're sharing the same data 6:06applications and policies between on 6:09premises infrastructure private Cloud 6:11public clouds and even out to the edge 6:13you're bringing critical data and 6:15applications into a consistent 6:17environment which means different 6:19business units can access the data they 6:21need without causing fragmentation or 6:23adding 6:24complexity essentially hybrid by Design 6:27means less Frank clouds more 6:31enterprise-wide alignment less time 6:33trying to access your data and more time 6:36actually using it to successfully build 6:39fine-tune integrate and deploy AI into 6:42your business now that we've defined the 6:44power of a hybrid by design it 6:46infrastructure let's answer some of the 6:48common questions around Ai and how does 6:51this topic even relate to hybrid Cloud 6:54first off can current technology provide 6:56a strong foundation for today's 6:58unparalleled AI capabil ities absolutely 7:01it can an intentionally designed hybrid 7:03Cloud architecture allows you to 7:05facilitate rapid scalability by tuning 7:07and deploying AI wherever your data 7:10resides this is a central step towards 7:12more efficient cost effective value 7:14creation second what does that 7:16infrastructure look like it looks like a 7:18hybrid Cloud environment but with a 7:20consistent platform and a consistent set 7:22of AI capabilities and governance across 7:25that full span it's a strategic coupling 7:28of public and private cloud scalability 7:30and on premises like security how can a 7:33hybrid by Design architecture help 7:35businesses gain the most value from AI 7:37well an intentionally designed hybrid 7:39Cloud architecture allows you to train 7:42modify and integrate AI capabilities 7:45into your business right alongside your 7:47most critical data that means that 7:50Solutions like Advanced automation 7:52real-time analysis and around the-clock 7:54customer engagement are right at your 7:56fingertips ready to be deployed when and 7:58where you need the most you're doing AI 8:01where your critical data is in all the 8:03places that you're doing business and 8:05where your customers are organizations 8:08need an integrated open and intentional 8:10hybrid Cloud architecture to enable AI 8:13deployment at scale hybrid by Design is 8:16all about deliberately building a 8:18cohesive it architecture that offers the 8:21combined benefit of multiple systems and 8:23services when your hyb Cloud 8:25architecture is strategically selected 8:28carefully structured and you leverage an 8:30AI vendor who meets you where you are 8:33your Enterprise can run AI workloads 8:35with the operational agility and 8:36scalability of the cloud underpinned by 8:39the robust security resilience and 8:41governance traditionally expected of on- 8:43premises 8:44infrastructure that is how you unlock 8:46the full value of AI