Implications of Generative Artificial Intelligence for the U.S. Intelligence Community: A Summary
Hello, I’m Ylli Bajraktari, CEO of the Special Competitive Studies Project. In this edition of our newsletter, SCSP’s Intelligence Panel provides an overview of the Intelligence memo of SCSP’s recent report, Generative AI: The Future of Innovation Power. The link to the full memo on Implications of Generative Artificial Intelligence for the U.S. Intelligence Community can be found here.
In the coming weeks, we will be sharing job opportunities to join the SCSP team. We will share the openings through our newsletter, social channels, and the career page on our website. Apply, share with your network, and continue to be a part of our important work. Thanks for following and believing in the mission.
Generative AI: The Future of Innovation Power Featured on SCSP’s NatSec Tech
Members of the SCSP team join Jeanne Meserve to summarize each memo from our fall report, “Generative AI: The Future of Innovation Power.”
In this sixth and final episode, SCSP’s Chip Usher summarize the memo on Implications of Generative Artificial Intelligence for the U.S. Intelligence Community.
The advent of Large Language Models (LLMs) and other forms of generative AI (GenAI) are transforming the business of intelligence. The Intelligence Community (IC) was early to recognize the potential of AI, but rapid advancements in GenAI tools necessitate even greater focus. This urgency is underscored by the fact that adversaries like the PRC and other state and non-state entities are already gaining a lead in leveraging these technologies. If the IC fails to capitalize on GenAI, it risks becoming ineffective or – worse – irrelevant to the U.S. national security mission. Yet, by embracing GenAI fully, there lies an opportunity to revolutionize intelligence workflows, drive a digital transformation, and harness the vast potential of open source intelligence (OSINT). Recognizing this, SCSP proposes actionable steps for the IC to seize upon the present moment and future possibilities.
Visualizing the Potential
From planning and direction to dissemination, GenAI can streamline and enhance various stages of the intelligence cycle—enabling faster anticipation of policymaker needs, automating data processing tasks, bolstering all-source analysis through AI-powered chatbots, and ensuring rapid and tailored information dissemination.
More fundamentally, GenAI stands as a catalyst for a profound cultural transformation within the IC. As the intelligence landscape undergoes what the National Academy of Sciences label a “revolution in intelligence affairs,” there's a push for the IC to shift from conventional secret intelligence methods towards an emphasis on OSINT. This change highlights the urgency for speed-to-insight and to leverage publicly available data. With the capabilities of GenAI, the IC can achieve this efficiently, economically, and with reduced risk to sensitive sources, better aligning with the pace set by industry and academia.
Actions the DNI Should Take Now
To fully capitalize on GenAI’s potential, the IC must quickly move beyond experimentation and limited pilot programs and begin deploying GenAI tools at scale. To lead this effort, SCSP recommends the Director of National Intelligence (DNI) insist that Agencies adhere to four critical principles regarding the IC’s use of AI:
Begin using GenAI tools immediately. GenAI will have a profound impact on IC work, so it's vital IC agencies start using these tools now to train the workforce and develop infrastructure and policies.
Become an “Agile Adopter.” This approach is more cost-effective and efficient than building state-of-the-art LLMs in-house. IC agencies should partner with a leading foundation model provider to acquire their LLM and either modify it to incorporate IC datasets and unique terminology or pair it with a smaller, IC-developed model.
Tackle privacy concerns up front. The Office of the Director of National Intelligence (ODNI) should work with IC agencies to address privacy concerns before using LLMs widely. This will help to prevent a varied approach to LLM use across the IC and reduce the risk of adversaries poisoning LLMs with private or protected intellectual property data.
Insist on IC-wide solutions wherever possible. AI expertise varies across the IC, and there is a risk of proliferation of inferior LLMs. The ODNI should use its authority to bring the IC together to acquire cutting-edge LLMs, build smaller LLM models, and set standards for AI use. Agencies should have flexibility to tailor their own models for specific purposes, but these should be the exception.
Additionally, SCSP recommends that the DNI appoint a senior IC lead for AI implementation to establish and lead a new IC Committee on AI Deployment and Use. The Committee should publish guidance on AI use, identify steps for faster acquisition and deployment, establish analytic tradecraft standards for GenAI use in finished intelligence, and design AI capabilities with allies.
Longer Term Efforts
While immediate actions are essential, they represent just a fraction of the broader strategic landscape. To maintain its edge and navigate the future with confidence, SCSP identifies three pivotal areas as cornerstones for the IC's enduring success in the long-term:
Increase Collection and Analysis on Foreign AI Capabilities. The DNI should task collectors to prioritize non-public insights into foreign AI plans. The DNI should also task the National Intelligence Council (NIC) to assemble an all-analysis expert network to assess foreign development and use of GenAI tools, including potential use of LLMs by adversaries. The NIC AI Red Team should present its findings to the White House and Congressional oversight committees by January 1, 2024.
Build an AI-Ready IC Workforce. The DNI should instruct the Assistant Director of National Intelligence/IC Human Capital to: (1) create universal “AI technical competence” standards, (2) build official GenAI tech talent career tracks, (3) revamp analytic incentives for GenAI use, and (4) leverage American expertise in GenAI for IC competitive advantage.
Reinvigorate the Open Source Mission. LLMs rely on OSINT data for training, necessitating an increase in IC usage of OSINT in all kinds. The ODNI should empower the new OSINT Executive to harmonize OSINT use across the enterprise and to identify successful programs and advocate for them to receive greater resources. The ODNI should also explore new alternative solutions, including the creation of an Open Source Agency or public-private partnership with industry.