人工智能是决定未来战争成败的新质战斗力,这一点在数字孪生战场实验室编写的《军事人工智能白皮书》中做了深入的分析。
报告围绕美军2014年提出的第三次抵消战略讲起,人工智能是该战略的核心关键技术。
然而,美军的AI到底发展到什么程度了?对于这个问题,《军事人工智能白皮书》紧紧围绕美军核心部门CDAO进行分析,重点研究了其2024年2月的“数据与人工智能峰会”(Data and AI Symposium)。
这次峰会透露了美军在AI发展上得到了民营企业的大力支持,已经处于“规模化”应用军事AI的阶段。
这场峰会有70名嘉宾参与发言或讨论,他们代表了美军的AI发展情况。
大家熟悉的应该有OpenAI、IBM、微软、亚马逊、甲骨文等知名企业,他们都是美军的重要合作方。
以OpenAI为例,它提供的ChatGPT、Sora等文生对话或视频的能力,目前毫无疑问处于全球领先地位,那么美军的AI应用,其竞争优势应该比较强大。
除了这些“如雷贯耳”的互联网企业,还有一些看似不知名的公司,例如Scale AI,这家公司的CEO直接排在美国国防部副部长凯瑟琳·希克斯(Kathleen Hicks)后面,位列第二位,排在了CDAO负责人马泰尔前面,可见美国国防部多么重视这家公司。
Scale AI的创始人和CEO是一位华人Alexandr Wang,他是从麻省理工学院辍学创业。他的父母是中国科学家,在美国核武器基地洛斯阿拉莫斯国家实验室(Los Alamos National Laboratory)工作,因此,自小就在数学和计算机上面有较高的天分。
从Scale AI公司的形成就可以看出,它的目的是为了“规模化”应用AI,实际上,这家公司确实是CDAO的铁杆支持方,它在高质量数据方面建立了较强的优势,客户包括OpenAI、Meta、微软、英伟达、丰田和哈佛商学院等。
除了Scale AI这样的创业公司,美军还把Lovelace AI、Robust AI、Geometric AI、Palantir、C3AI等大家不太熟悉的中小微民营企业纳入其中,可见是真的“不拘一格降人才”。
人们可以从这次峰会了解到美军的AI进展,以及到底在解决什么问题。
建议大家查看这次数据与人工智能峰会的议程,从中有可能了解更多美军的AI相关实力,结合到《军事人工智能白皮书》,读者不难洞悉其发展方向。
议程如下:
Monday, February 19
1600-1900
Attendee & Exhibitor* Badge Pick-Up
*Exhibit Hall will not be secured overnight on 2/19, please do not keep valuables in your exhibit area.
Tuesday, February 20 – Where We Are
0830-1230 Attendee Check-In
1230-1245 Security Briefing, Announcements, and Opening Remarks
1245-1345 Kickoff Keynote: The State of AI and the CDAO, Dr. Craig Martell, Chief Digital & AI Officer
1345-1430
DoD GenAI Use Cases and Acceptability Criterias
Don’t FAR on my OTA
Test & Evaluation of Evolving Systems
1430-1500 Exhibit Hall Networking Break
1500-1545
Developing a Culture of Experimentation
Applications to Enable Decision Advantage
DAGR, SHIELD, & the DoD Responsible AI Toolkit Workshop: Federated Computational Governance for Data Mesh
1545-1630
Assuring LLMs for DoD Applications
Boosting DoD Performance with Pulse Data Analytics
Firetalks: Learning By Doing – Partners in Pathfinding to Operation Ready Data & AI
Workshop: The Forgotten Requisite Resource – Compute, Platforms, and Sandboxes
1700-1900 All-Hands Networking Social
Wednesday, February 21 – What the DoD Needs
0730-0830 Attendee Check-In
0830-0835 Security Briefing and Announcements
0835-0905 Fireside Chat: AI Strategy and Implementation
Ms. Margaret Palmieri, Principal Deputy Chief Digital & AI Officer
Mr. John Turner, Acting Deputy Chief Digital and Artificial Intelligence Officer (DCDAO) for Policy
Mr. Andrew Peppler, Policy Analyst, Chief Digital and Artificial Intelligence Office
0905-0935 Keynote: The DoD’s Technical Needs
Dr. William Streilein, Chief Technology Officer, CDAO
0935-1000 Exhibit Hall Networking Break
1000-1045 Achieving Decision Advantage as a Combat Support Agency
(1000 – 1100) Opening Remarks for LLM Maturity Models Workshops
(1000-1015)
Creating a Fluid Workforce for a Digital Future
Advancing Data Exchange/API Efforts to Increase Access & Sharing
1045-1130 LLM Maturity Models
(1015-1130) Workflow Integration of AI Working With Small Business
1130-1215 Global Conversation on the Responsible Use of AI in Defense
(1130 – 1230) Workflow/System Integration of Maturity and Community Discussion
(1130-1215)
How Data & Analytic Products are Changing
How the DoD Supports Incident Response
Risk Centric, Agile Cyber Assurance
1215-1345 Lunch On Own & Exhibit Hall Networking Break
1345-1415 Keynote: The Global AI Contest
Honorable Kathleen Hicks, Deputy Secretary of Defense
1415-1500
The Data Integration Layer: CJADC2 Data Interoperability and Accessibility
Workflow/System Integration of Maturity
(1430-1455)
Solving the Right Problems: Why User Research Matters
Improving Outcomes through Enabling Dev Access to Mission Data
1500-1530 Exhibit Hall Networking Break
1530-1615
Mission: Democratize Data & Operationalize AI
Maturity Model Validation and Use Case Exploration
(1530-1645)
Adversarial AI: Bridging Research With Operations
LLMs and Cybersecurity: Practical Examples and a Look Ahead
1615-1700
MILDEP CDO Perspective
Closing Remarks for LLM Maturity Model
(1645-1700)
Learning from User Decision Making and Behaviors – Instrumentation and Measuring of AI-enabled Systems
1700-1705 Closing Remarks
Dr. Craig Martell, Chief Digital & AI Officer
Thursday, February 22 – Where We Are Headed
0800-0830 Attendee Check-In
0830-0845 Security Briefing and Announcements
0845-0945 LLM Potential – A Discussion
Moderator: Dr. Craig Martell, Chief Digital & AI Officer
Panelists: Mr. Alexandr Wang, CEO, ScaleAI
Dr. Gary Marcus, Professor Emeritus of Psychology and Neural Science, New York University
Dr. Andrew Moore, Founder, Lovelace AI
Mr. Shyam Sankar, Chief Technology Officer and Executive Vice President, Palantir Technologies
0945-1015 Exhibit Hall Networking Break
1015-1100
The Digital Ecosystem and Project Linchpin
Research Challenges for Large Pre-Trained Models
Recruiting, Retaining, and Employing Digital Talent in Global Defense (U.S. and Allied Partner Nations)
Responsible AI Tools Deep Dive
1100-1145 Ethical Use of LLMs
Accelerating CJADC2 Through Experimentation
Recruiting and Retaining AI Talent – Learning from Industry
1145-1300 Lunch On Own & Exhibit Hall Networking Break
1300-1345
The Future of LLMs
Operationalizing AI Ethics in the Military and Intelligence Domains
1345-1430
The Future of AI
Empowering Self-Serve Analytics in the DoD
1430-1500 Exhibit Hall Networking Break
1500-1530 TF Lima – The Way Forward (Goals and Progress)
1530-1545 Closing Remarks, Dr. Craig Martell, Chief Digital & AI Officer
Friday, February 23 – Classified Sessions
0830-0835 Welcome Remarks, Dr. William Streilein, Chief Technology Officer, CDAO
0835-0915 Context Keynote, Mr. Lincoln Davidson, Strategy & Policy Analyst, CDAO
0915-1000 CAMELIA Workshop Update (ODNI)
1000-1045 Break
1045-1130 Project Maven Partnership, Vice Admiral Frank Whitworth, Director, National Geospatial-Intelligence Agency
1130-1215 NSA AI Security Center, Mr. Vinh Nguyen, NSA
1215-1330 Lunch
1330-1430 Red Capability Sessions, NGIC – AI Team
1430-1530 Blue Capability Sessions
1530-1545 The Way Forward – Closing Remarks
Dr. William Streilein, CDAO
Mr. Shane Partlow, CDAO
已经购买《军事人工智能白皮书》的客户,可以获得以下会议资料:
1 The State of AI and the CDAO
2 Applications for Decision Advantage
3 LLM Maturity Model For Ai Risk Management In The DOD
4 Large Language Model (LLM) Maturity Model
5 Maturity Model Opening Remarks
评论