New Evolutions in GenerativeAI Present Significant Opportunity

By Dave DeWalt, CEO, NightDragon

The advent of our new age of artificial intelligence has captured the world’s attention over the past few months. With immense possibilities around content creation, automation, analysis and more, artificial intelligence has become the topic of nearly every conversation, from the kitchen table to the boardroom.

But are we already moving past the current generation of Generative AI as we know it? In some ways, yes. Where ChatGPT and other types of popular Generative AI technologies can fall short is that they are typically prompt-based. What that means is that they require a human question to be asked in order to provide an answer (though it is exceptionally good at that).

We’re seeing advancements emerging already to close that gap and drive AI forward in a meaningful way. Regenerative AI is a new form of Generative AI that automatically regenerates in real time, without the need for prompting. While previous generations of Generative AI needed a question to spur an answer, this new development in AI just skips right to the answer that you need, even before the question is asked. As a result, it regenerates what content it produces as new information evolves, all on its own without prompting.

There are many use cases where this type of technology can be helpful. One area already in use by a company called Dataminr (who developed the technology) is around discovering events, risks and threats with unparalleled speed, scope and accuracy. When an event occurs, how can we use AI to ingest millions of data points on events and update us with a summary on the latest information and how that event may impact us? This is possible with the use of Regenerative AI.

This technology was put to the test in the midst of the recent Francis Scott Key Bridge collapse in Baltimore. Dataminr alerted on the catastrophe 74 minutes before major news networks (as soon as the bridge was shut down to cars), and then used Regenerative AI to automatically update without prompting to provide the latest information on the scope and scale of the incident as more details emerged. Similarly, it was also used to track the recent attack on Israel from Iran in a similar way.

These incidents are multidimensional events that evolve rapidly as new information emerges, such as the name and origin of the ship in the Baltimore bridge collapse or details on the rescue efforts for victims. This requires constant updating on the status of the event. In the past, this would have required prompting AI for each update, but Regenerative AI can update automatically.

Where can this go in the future? One exciting potential for Regenerative AI is to take it beyond just the capabilities of what information is available today, but to predict what could happen in the future. For instance, in our example of it being used to synthesize key developments in world events, could it also be used to predict what could happen in the future, versus just what is happening in the present? Those capabilities are incredibly compelling and in development right now.

AI is an incredibly powerful technology, but it is clear that we have not seen the end of innovation coming out of this sector. We are only at the beginning of our new age of AI with many exciting things to come.