Generative AI has the potential to be transformative across the entire organization—including inside IT’s own walls.
IT is at a fascinating inflection point. For years, they’ve been tasked with both running and modernizing the infrastructure that underpins businesses—a challenge not for the faint of heart. Whether they quickly embraced the cloud for its speed and agility, or followed a more measured path to modernization, they’ve likely grappled with significant disruption in the past decade. Now they face a new seismic shift: the emergence of generative AI.
Many IT organizations are understandably nervous. As internal users adopt publicly available solutions, the need for centralized security controls and governance has become increasingly urgent. And as we’re still in the early innings of generative AI, knowing how best to bring it to organizations while blocking and tackling its risks is very much a work in progress.
But while IT organizations work to solve these challenges, they shouldn’t miss a larger opportunity for themselves: generative AI has the potential to transform IT as well. In the face of increasing demand and skills shortages, IT organizations willing to embrace generative AI have the potential to be among the first in the enterprise to reap its transformative benefits.
How Generative AI Can Democratize the Data Center
Giving team members the ability to query data using natural language prompts, or easily generate code outside of immediate domain expertise, has the potential to simplify and accelerate IT operations. This can have enormous ramifications, lowering barriers to entry to many types of modern practices.
Consider DevOps. Practitioners often rely on code snippets to perform routine tasks. With generative AI, instead of searching the internet for the right code, they can generate it for their specific need, often faster and at a higher quality. By embracing automation and generative AI, code developers will see fewer errors, and they can use the tools to help explain errors or create unit tests. The same is true of documentation—instead of traversing the web, they can prompt a chatbot and get exactly what they’re looking for.
Granted, in these examples, there’s no guarantee the right information will be available, or error-free—but this is also true of internet searches. But in cases where generative AI can reduce the time and effort getting to the right solutions, the potential upsides are significant. Furthermore, the time that would have been spent creating the solution can now be spent on testing it and reviewing the code.
If you think about it, the emergence of infrastructure-as-code has enabled a level of automation within data centers that was unimaginable a decade ago. Common infrastructure scripts are readily available and can be pushed to servers at scale, lowering the bar on managing an IT environment. All you need to do is find the right scripts. Enter generative AI.
The GPS on GPT: Where Do We Go From Here?
With all that said, generative AI will not cover every single use case, and it’s not without its challenges and risks. For starters, there’s only so much a generative AI model trained on public data will know. Although tools like ChatGPT are surprising versed on publicly available developer data, they don’t cover every user scenario or edge case.
There’s also broad concern about the accuracy of generative AI outputs and the overall risk of prompting them with sensitive or proprietary data. According to a recent Harris Poll, 51% of respondents said their top concern around implementing generative AI was quality and control, and 49% said their top concern was safety and security risks.
For these reasons, many organizations are looking beyond merely adopting publicly-available generative AI solutions and building their own foundation models in house—or adopting pre-existing models and training them with proprietary data. This practice lets them train the model on their own proprietary data, which allows for more refined results with less risk to security or IP leakage. Organizations gain more control over all variables, which can result in a more customized and useful model for internal customers, increased security and intellectual property protection, and helps them set better guardrails overall.
The Generative AI Era of IT Starts Now
In many ways, discussions around generative AI are a natural progression of conversations we’ve had for years around the cloud and IT modernization. Organizations that laid the groundwork for cloud adoption and on-premises automation may be readier than they know for seamless integration with generative AI’s capabilities. If they have standardized cloud and automation processes, they may be able to rapidly adopt generative AI with just a single API call.
This new era holds the potential to significantly reduce barriers to entry in countless places. Where the road leads is still unclear, but one thing we know for certain: one of its most exciting destinations lies inside IT.
Learn how generative AI and automation lower the bar for data center management in our latest podcast, The Great Equalizer: Gen AI and AI Transforming the Data Center and learn how to bring generative AI to your organization.
Read the full article here
