Booz Allen Hamilton Inc., the McLean-based technology and management consulting giant, has launched a suite of free artificial intelligence tools it says will protect sensitive data better than other leading AI platforms.
The platform, called aiSSEMBLE Baseline and available on GitHub, is designed to simplify the building and deployment of AI-based systems. It emerged from client work Booz Allen (NYSE: BAH) has completed on its closed-source platform to bring AI technologies to government, academic, commercial and nonprofit worlds.
Bryan Castle, director of software engineering for the platform, said aiSSEMBLE Baseline keeps the data users input private unlike other free AI offerings, which often use this data for training of their broader public offering. That was a crucial requirement for many Booz customers, particularly its public sector clients.
"They very often need to train and build their own models using their own data sets that are very sometimes sensitive and custom and they can't just send that off to a public cloud," Castle told me. "Very often, we're being asked to process their sensitive data and then use that sensitive data to train custom AI models and then deploy those AI models into production so we can do things like inferencing."
Castle told me aiSSEMBLE Baseline isn't a typical software product but rather a tool kit businesses or organizations can use that can be tailored for specific applications such as those that rely on generative AI or image-based creations. He said the company views AI as "the critical technology of our time."
"Applications of AI are popping up all the time, but there's often a consistent set of patterns that we can identify across all of those, and aiSSEMBLE really takes those patterns and implements them as a set of reusable blueprints so we can bring them across all of our different client efforts," he said. "Don't think of it as something that you just install off the shelf and use, but think of it as something that's really used to build other things, and in this case, fit-for-purpose applications to meet our government clients' very specific and nuanced needs."
The National Oceanic and Atmospheric Administration (NOAA) is one of Booz Allen's first customers to evaluate a use case for aiSSEMBLE Baseline. The federal agency is inputting weather data into aiSSEMBLE Baseline's AI and machine learning algorithms to see if the technology can improve employee productivity.
By making the product free and open source, Castle said he's optimistic others will be able to uncover additional potential benefits or situations that can use the technology.
"We think that by making aiSSEMBLE open source, it simplifies the IP restrictions and the ability for our government clients to be able to adopt it even independent of Booz Allen," he said.
Booz Allen now has 2,300 people working on AI services and solutions, a figure that has jumped 25% annualized since the close of the company's 2023 fiscal year in March. The company, which employs nearly 34,000 globally, posted $9.3 billion in revenue during its fiscal year ended that ended in March 2023.
Castle said AI developments like the new platform serve as "a core part" of Booz Allen's business model and it's "one of the biggest drivers" of growth that the company has seen over the last several years.
"We expect that it will continue to be one of the biggest drivers of growth; there's almost not a single contract that we encounter that is not asking for some kind of AI in some way, shape or form," he said. "It takes a lot of different formats, but almost everyone is looking to have AI transform their business."
