I think it is safe to say that large language models and generative artificial intelligence, like ChatGPT, have taken the worlds of technology and work by storm. The unique capabilities of ChatGPT present an opportunity to enhance productivity across a wide number of functions.
New research out of MIT supports this thought:
“The ChatGPT using group was 37% faster at completing tasks (17 minutes to complete vs. 27 minutes) with roughly similar grades (level of quality), and as the workers repeated their tasks for improvement the ChatGPT groups quality went up significantly faster. In other words, ChatGPT did make work speedier with no sacrifice in quality and then made it easier to “improve work quickly” using the tool.”
Knowing that ChatGPT and other AI solutions possess great potential to help most of us do our jobs more efficiently, how can this technology solution and others like it support the intelligence function? One of the areas of particular interest to me is using Chat GPT to develop arguments. As you know, developing arguments involves reviewing a mass of information to build a case that includes identifying hypotheses, claims, subclaims, evidence and assumptions.
I decided to put my query to the test. I asked CHATGPT to:
Review the last 10 years of threat assessments from the Office of the Director for National Intelligence. Develop arguments on China’s space capabilities over that time span. [hypothesis] [claim] [subclaim] [evidence] [assumptions]
ChatGPT provided this response (Note: This response hasn’t been edited/proofed by a human.):
Hypothesis: Over the last 10 years, the Office of the Director for National Intelligence has identified China’s space capabilities as a growing threat to US national security.
Claim: The assessments from the ODNI over the past decade have emphasized China’s rapid development of space capabilities, which have the potential to challenge US military dominance and threaten critical infrastructure.
Subclaim 1: China has made significant progress in developing advanced space technology, including satellites, rockets, and manned spacecraft.
Evidence: The ODNI assessments have highlighted China’s successful missions to the Moon, Mars, and other planets, as well as the country’s plans to build a space station by 2022. Additionally, China has conducted multiple anti-satellite tests and has developed a range of satellite-based capabilities, including navigation, remote sensing, and communication.
Assumptions: China’s rapid development of space capabilities is driven by a desire to achieve parity with the US and other space-faring nations, as well as to support its military and economic goals.
Subclaim 2: China’s space capabilities pose a significant threat to US national security, including the potential to disrupt critical infrastructure and military communications.
Evidence: The ODNI assessments have noted China’s development of anti-satellite weapons and other technologies that could be used to interfere with or disable US satellites. Such disruptions could have far-reaching consequences, including damage to the US military’s ability to communicate and coordinate operations.
Assumptions: China’s space capabilities are part of a broader effort to challenge US military dominance and exert greater influence in global affairs. China may be willing to use its space capabilities to achieve strategic goals, even if doing so risks conflict with the US and its allies.
Overall, the ODNI assessments over the past 10 years suggest that China’s space capabilities have rapidly advanced and could pose a significant threat to US national security. While the full extent of this threat remains uncertain, it is clear that the US must take steps to monitor and mitigate potential risks posed by China’s space program.
So, now I ask you: Is this a complete argument? Is it valid? What biases might exist in the argument? What are your thoughts?
Personally, I believe the benefits of using ChatGPT to answer this sample query are:
- ChatGPT reviewed 10 years worth of ODNI threat assessments in 10 seconds, developed a hypothesis and started to layout the other argument components per my original query
- ChatGPT provided a means to graphically represent this argument if I decide to do so
- ChatGPT identified assumptions in its argument that could lead to the need for additional collection management to turn the assumptions into evidence
I assert that ChatGPT and other AI solutions similar to it are going to impact the missions, strategies and workforce roles and responsibilities in a wide swath of organizations in myriad ways in the future (in both the public and private sectors), including those in the Intelligence Community. Much of this impact is obviously still to be determined and the potential is truly mind blowing.
The key is to get ahead of (or at least try to keep pace with) the Chat GPT and similar technology solutions is to prepare workforces to understand them (who, what, where, when, why and how) to improve productivity and effectiveness for mission success.
What are your impressions of ChatGPT and its impact on the IC? Let’s start a conversation.
Note: FedLearn is currently building an introductory ChatGPT course for its Intelligence catalog to be released in 2023.
Dr. J. Keith Dunbar
Founder and Chief Executive Officer