Digital transformation is on everyone’s minds across the federal government. For the intelligence community more specifically, agencies want to stay on top of artificial intelligence, machine learning and biometrics development.
For John Beieler, the director of Science and Technology for the Office of the Director of National Intelligence, that means making sure agencies such as the CIA, National Security Agency and Energy and State departments invest their dollars in S&T capabilities. And from there, close any gaps from one agency’s mission to the next – and scan the horizon for future applications.
“So this ends up applying to lots of technological areas like artificial intelligence, machine learning, but also tons of other areas to say, ‘Hey, agency x is doing a great job at their particular mission outcome, but what is the community doing?’ We need to consider that each specific agency might not think about when executing their own mission,” Beieler said on Federal Monthly Insights – Digital Transformation.
The ODNI Science and Technology directorate considers 19 technology domains currently, from AI and machine learning to the biological sciences, quantum information science, quantum sensing, precision navigation and timing, and high performance computing – to name a few. The IC is also having a larger conversation about data use, which includes biometrics.
“The biometrics conversation as part of this kind of larger push to how we consider our data and how we posture ourselves to do the fundamental mission, the intelligence community, which is collect and analyze information to provide insights to policymakers within an actionable timeframe,” he said on Federal Drive with Tom Temin. “So that all gets kind of wrapped up in this, how do we make use of the data that we have to do our job better, faster, and more accurately.”
In the ODNI context, digital transformation is behind the AIM Initiative, or augmenting intelligence using machines. Beieler said the initiative focuses on how the IC deploys “triple A technologies” of augmentation, artificial intelligence and automation.
“Everyone gets very focused on machine learning. But there’s a whole host of solutions that aren’t the kind of cutting edge machine learning research or deep neural networks or things like that. So I always joke that a lot of times, we have analysts working with a set of spreadsheets, and they don’t really need a deep neural network, they need a Python script or something like that, or just some simple automation techniques,” he said. “The goal is to apply these technologies and techniques to the large amount of information which can be surfaced to our analysts, and then draw conclusions that policymakers need for informed decision-making.”
Shareability across the IC is key – nobody wants to duplicate efforts, he said. But duplicate investment in nascent technologies may be acceptable if it’s not yet known what the best use for those technologies is, and different agencies could have different applications for them.
The pandemic has disrupted traditional intelligence gathering, although Beieler declined to comment on specific agency missions to that effect. What he did say was how externally developed commercial technologies, such as facial recognition, can be harvested for agency use.
“So what we are really focused on is how do we tap into a lot of those things that are happening out in the open and adapt them, or modify them to what we need to do our mission, as well as, if we’re building something completely on our own, how do we ensure that we’re using the most state of the art techniques and the most state of the art technologies, even if we’re building something on our own?” he said.