This is a key time in the evolution of artificial intelligence (AI). In the past decade, I’ve had a front-row seat to the countless false dawns of AI. This time, things are different.

‍

It’s clear to see why: today, more than ever before, we have the ability to manage and store vast amounts of data cheaply and efficiently. We have the raw materials needed to build, deploy, and train powerful compute models. Advances in underlying infrastructure, like highly capable Graphic Processing Units (GPUs), have provided the processing power needed to support the requisite training. We have advanced algorithms that can take advantage of the compute infrastructure and vast troves of data to generate rapid and deep analyses that far exceed the capabilities of previous generations of AI.

‍

Today, we’re realizing the groundbreaking possibility that AI had always promised to be. And at IQT, we are matching those possibilities up with our greatest national security needs. From information gathering and automated recognition to advanced quantum computing and assured positioning, timing, and targeting, AI is opening new doors to security innovation.

‍

As an example, among the most daunting challenges we face right now is trying to make sense of the gargantuan amounts of information and open source data we’re faced with.

‍

The challenge isn’t “information overload,” it’s “filter failure.” We need to become better and faster at cutting through the chaff to get to the information that matters. Generative AI “copilots,” or AI-powered software assistants, can be invaluable to the rapid identification of key information and patterns, in addition to automating data-gathering and dissemination tasks. They can also boost existing capabilities in areas such as natural language processing, enhance real-time translation, and help to understand foreign-language content.

‍

And it’s not just data ingest and synthesis. On the visual recognition front, AI is a force multiplier. Today’s computer vision models are exponentially more powerful than ever before and represent a huge change in capabilities from their predecessors. They support nearly any vision task while performing orders of magnitude higher speed and accuracy.

‍

AI promises to enhance the U.S. Intelligence Community’s ability to identify valuable sources of intelligence. It can accelerate the accurate prediction of preferences and behaviors after being trained on past actions and behavior patterns. This ability to predict future behavior can be leveraged to focus collection and analysis.

‍

All of the above frees our most valuable resource, people, to focus on the more pressing tasks at hand.

‍

Despite its benefits, it’s important to remember that generative AI (GenAI) is a digital double-edged sword; these same capabilities can be used against us by state and non-state actors alike, which is why it is critical we invest in tech that can help to combat the work of adversarial AI. Cracking this ‘black box’ of AI decision-making is an important focus area of IQT. If we can understand how sophisticated technologies arrive at their conclusions and decide upon a course of action, we can more effectively prepare and respond.

‍

We’re seeing the very beginning of a technology which will surely bring transformational changes to the way we live and work. Already we are realizing the ways in which it can be leveraged for both good and bad and are trying to discern the best ways to combat the latter without interrupting the former. At IQT, we are singularly focused on investing in transformational technologies—in AI and elsewhere—which will maintain our competitive advantage and secure our national interest.

‍

I see no false dawn with this iteration of AI technology and am optimistic in our country’s ability to make good on the promise of a safe, secure, AI-enabled future.