The intent of this article is to encourage a reflection about the inexorable use of technologies arising from Artificial Intelligence (AI) for the benefit of Intelligence, the classic activity practiced by Intelligence Agencies.
The motivation to write these words is due to the publication of the excellent article “The US intelligence community is embracing generative AI”, authored by Mr. Frank Konkel.1 At its core, it cites positions of the Head of Innovation for Artificial Intelligence, no less than the Central Intelligence Agency (CIA), which recently took place at the Amazon Web Services Summit, in Washington, D.C.
According to the Konkel’s article, analysts from the Intelligence Community in the US would already be using Generative AI in classified work environments, probably considered as confidential, to assist in data collection, writing, ideation, brainstorming and generation of counterarguments, given the pre-existing capabilities of translation and transcription of human language and data processing, that I’d reference it as Big Data.
Also, in parallel, Amazon Web Services introduces generative AI as:2
Type of AI that can create new content and ideas, including conversations, stories, images, videos, and music. AI technologies attempt to mimic human intelligence in non-traditional computing tasks such as image recognition, natural language processing (NLP), and translation. Generative AI is the next step in artificial intelligence…
…
Generative AI algorithms can explore and analyze complex data in new ways. Thus, researchers can discover new trends and patterns that would otherwise not be apparent. These algorithms can summarize content, outline various solution paths, brainstorm ideas, and create detailed documentation from research notes. That’s why generative AI dramatically enhances research and innovation. (Emphasis added – Translated from Portuguese)
According to the CIA Officer, AI helps Intelligence Analysts sift through large amounts of data in open sources (Open Source Intelligence – OSINT) to extract insights, in addition to classifying them, which in this gigantic haystack, AI can help find the needle and, thus, subsidize decision-makers, including those of a political nature.
As we know, the role of Intelligence professionals, among them I highlight that of Analyst, is to produce KNOWLEDGE, the noble product of Intelligence, intended to advise a decision-maker.
The sources of intelligence data include OSINT, as already mentioned, IMINT, SIGINT, MASINT3, and HUMINT. The latter, Human Intelligence, for which I highlight special attention in the Intelligence Activity, due to the human factor, which I point out in the book Contra & Inteligência 4.0 (p. 86) as being the
source coming from people who transmit in a deliberate, indirect, inadvertent, unconscious way, without realizing the act of passing information to a third person, or under a certain condition of pressure, the knowledge they have on a subject that is the object of interest to an intelligence agency.
Undoubtedly, these data sources can feed and help the algorithms of a generative AI and, thus, contribute to the work of analysts to produce Intelligence Knowledge. However, a possible first question would be: Would Generative AI be another source of data collection, or would it be a data evaluation tool, one of the essential stages or phases of an Intelligence Analyst’s job?
As mentioned, there is the intent to encourage possible reflection on the subject, in the face of a political and socioeconomic environment, in which we experience a framework or landscape of Information Disorder4 (Baptistella, p. 213). Besides, the terms Information Warfare and Cognitive Warfare have become increasingly evident and provocative, particularly for analysts.
Furthermore, I mention in Contra & Inteligência 4.0 (p. 98) that Cognitive Bias5 is one of the biggest pitfalls in Intelligence Activity because it involves a pattern of systematic errors in thinking that can distort the judgment of an individual or a group of people during the reasoning process.
Therefore, the suggestion for reflection to be formulated would be: How can Intelligence Analysts avoid possible or eventual interference of cognitive biases when working in the process of collect, evaluate and analyze data, using Generative AI?
Not by chance, Konkle’s article mentions that “Countless reports showcase so-called “hallucinations” — or inaccurate answers — spit out by generative AI software. In national security settings, AI hallucinations could have catastrophic consequences. Senior intelligence officials recognize the technology’s potential but must responsibly weigh its risks.”
Luis F. Baptistella
Bravus Consultoria
1 Available at: https://www.nextgov.com/artificial-intelligence/2024/07/us-intelligence-community-embracing-generative-ai/397849/. The article was published on 07/03/2024. Accessed in: 07/07/2024.
2 Available at: https://aws.amazon.com/pt/what-is/generative-ai/ . Accessed in: 07/07/2024.
3 IMINT – Imagery Intelligence; SIGINT – Signals Intelligence; MASINT – Measurement and Signature Intelligence.
4 See https://www.researchgate.net/figure/Categorization-of-Information-Disorder_fig1_352887207, related to Misinformation, Disinformation and Malinformation.
5 I recommend reading Spies, Lies and Algorithms by Amy B. Zegart (2022) which addresses this topic in depth.