澳大利亚战略政策研究所：2024年情报分析的未来美澳人工智能与人类人机协同项目（英文版）（42页）.pdf

1、The Future ofIntelligence Analysis:U.S.-Australia Projecton AI and HumanMachine TeamingSeptember 2024 Page 2 Contents Executive Summary 3 Scope Note 6 Introduction 7 The Intelligence Analysis Mission and Expectations of Generative AI 9 AI Today:Thinking About Applications Across the Analytic Workflo

2、w 14 Looking Ahead:The Coming Wave of AI Advancements 22 Recommended Actions 27 Appendix A:Factors That Influence Generative AI Model Performance 36 Appendix B:Differing Perspectives on AIs Potential 38 Contributors 40 Page 3 Executive Summary Rapid advances in the development of artificial intellig

3、ence(AI)technologies since late 2022,particularly the deployment of Generative AI(GenAI)chatbots powered by large language models(LLMs),have demonstrated the potential for AI to revolutionize how states conduct intelligence work.AI technologies are very likely to continue to rapidly advance given th

4、e large amount of investment from the private sector and nation states,with some experts predicting we will see the advent of artificial general intelligence(AGI)a type of AI that achieves,or surpasses,human-level capacity for learning,perception,and cognitive flexibility by the end of this decade.1

5、 Even if this ambitious goal is not fully met,the LLMs available within the next three years will probably far surpass the capabilities of systems we use today and will be able to solve complex problems,take action to collect and sort data,and deliver well-reasoned assessments at scale and at speed.

6、The effects of AI likely will be felt at all levels of the intelligence enterprise,including in collection,but the arena that we assess will see the earliest impact will be on the all-source analytic mission because of AIs ability to quickly process large volumes of data and GenAIs ability to produc