Takeaways: How intelligence agencies' are cautiously embracing generative AI
U.S. intelligence agencies are scrambling to embrace the AI revolution, convinced they’ll otherwise be smothered in data as sensor-generated surveillance tech further blankets the planet
ARLINGTON, Virginia (AP) — U.S. intelligence agencies are scrambling to embrace the AI revolution, convinced they’ll otherwise be smothered in data as sensor-generated surveillance tech further blankets the planet. They also need to keep pace with competitors, who are already using AI to seed social media platforms with deepfakes.
But the tech is young and brittle, and officials are acutely aware that generative AI is anything but tailor-made for a trade steeped in danger and deception.
Years before OpenAI's ChatGPT set off the current generative AI marketing frenzy, U.S. intelligence and defense officials were experimenting with the technology. One contractor, Rhombus Power, used it to uncover fentanyl trafficking in China in 2019 at rates far exceeding human-only analysis. Rhombus would later predict Russia’s full-scale invasion of Ukraine four months in advance with 80% certainty.
