Data is the new oil, but how do we drill for it? This is the question that must be answered today, and it is something we are seeing play out at the Olympics. In fact, Kaveh Mehrabi from the Intl. Olympics Committee, says the challenge we have is the volume of information.
“As much as information is good, it can also become overwhelming. It can be difficult to understand and be confident where is the source of truth,” he says.
This is one of the reasons the Intl. Olympics Committee and Intel collaborated to create a custom-designed AI Chatbot—Athlete365—to help roughly 11,000 competitors with varying languages and cultures navigate the venue and comply with the rules and regulations.
Powered by Intel Gaudi and Xeon processors, the chatbot delivers on-demand information to Olympic athletes during their stay in the Olympic village to help them focus on training and competing. The GenAI RAG (Retrieval Augmented Generation) solution can handle inquiries and interactions while delivering information.
The Rise of AI
AI (artificial intelligence) is rapidly changing how businesses operate. For one, we are seeing a rise in demand for AI. In fact, Intel suggests by 2026, it expects 80% of enterprises to have adopted gen AI in some form.
“We also realize that demand for AI and the demand for compute that it is going to drive is putting a new spotlight on availability, performance, cost, energy consumption and efficiency, and of course, data security,” says Justin Hotard, executive vice president and general manager of data center, AI business, Intel.
Developers also recognize the shift that is happening, as they are also really focused on the consumption of data.
“There is data coming from everywhere,” says Steven Huels, VP and GM, AI, RedHat. “There is more than any individual can process. For people to make meaningful use out of it, whether it is individuals like athletes or the general public or enterprises, you have to really focus on the usability and consumption of that information.”
And this is precisely what is happening at the Olympics this summer. But, of course, the Olympics are only one example. Deploying gen AI in any industry poses a number of challenges.
“What is relevant for the enterprise is data,” says Bill Pearson, VP software solutions, Intel. “Enterprises have tons of data often proprietary data, historical data, data they need to put to use. When you look LLMs (large language models), they are a great innovation and invention, but they don’t have access to that enterprise data.”
To help, the GenAI RAG solution will allow Intel to marry those two things together: the power of LLMs and the specificity and accuracy of the enterprise data. Built on the OPEA (Open Platform for Enterprise AI), the solution is flexible and customizable to an enterprise’s needs.
At the end of the day, the quality of the data matters. Oftentimes, data is based on popularity (think search engines, recommendations, and social media, as a few examples). But Rob Clark, president and chief technology officer, Seekr, suggests we need to focus on credibility and reliability when we are training the models, as that has a real downstream impact on business.
Some important questions to ask are: Can you rely on that data? Can you rely on that model? When we talk about responsibility, it is about the understanding. After all, Clark says, “You must build on a foundation that can be trusted. You must have the tools to validate as well.”
Want to tweet about this article? Use hashtags #IoT #sustainability #AI #5G #cloud #edge #futureofwork #digitaltransformation #green #ecosystem #environmental #circularworld #data #Olympics