Artificial intelligence is now beginning to take on the work of intelligence analysts. According to a report, the US Central Intelligence Agency (CIA) is deploying AI to analyse intelligence gathered from its human spies. The technology will help the agency analyse the plans, capabilities, and intentions of foreign nations and create a holistic assessment of issues for policymakers to decide the further course of action. The application of this technology is expected to make the process quicker and more rigorous.
The agency has already created its first-ever autonomous intelligence report, and Deputy Director Michael Ellis has said that its use will continue to grow.
According to a Politico report, Ellis stated that the agency will build a classified version of generative AI across its analytic platforms to assist analysts with basic tasks such as drafting key judgments, testing conclusions, and identifying trends in information.
While AI will handle much of the analysis, the agency has clarified that final decision-making will remain in human hands.
The US is already deploying AI in its war with Iran. In March, Brad Cooper, head of the US Central Command (CENTCOM), said that AI is helping US soldiers process large volumes of data.
In the ongoing US-Iran war, AI has helped US forces analyse vast amounts of data in seconds and make faster, more informed decisions than their adversaries.
Tensions with Anthropic
This development comes at a time when the US Department of War is engaged in a legal battle with AI company Anthropic. The US government has blacklisted the company and designated it as a supply chain risk after it attempted to limit the use of its models allegedly for lethal attacks and mass surveillance.
Ellis said that the agency will “not let private companies dictate how and when the CIA will make lawful use of their technologies.”
Ellis added that the agency is deploying AI to maintain an edge over China, which he believes is now close to the US in terms of technological innovation.
The agency has already tested around 300 AI projects over the past year, including those focused on processing large datasets and language translation.
