北京大学 · 国际关系学院 · 全球风险政治分析实验室Peking University · School of International Studies · PKU Analytics Lab for Global Risk Politics 承办实验室官网 ↗EN中文
「战略智能」挑战赛
Strategic Intelligence Challenge
主办 / 北京大学国际关系学院承办 / 全球风险政治分析实验室
U
建院 30 周年 · 特别活动

战略智能」挑战赛

AI 赋能地缘政治风险分析与专精特新全球化决策推演

数据 · 洞察 · 决策

距一号公告报名截止

报名倒计时

00
00
00
00
赛事周期约 4 个月 · 15 周本 / 硕 / 博 全面开放

大赛简介

在生成式人工智能引领的智能革命与全球「百年未有之大变局」历史性交汇的当下,2026 年正值北京大学国际关系学院建院 30 周年。学院依托深厚学术底蕴,将沉淀三十年的宏观地缘理论与国别经验,升华为护航中国企业全球化生存的「数智基础设施」。

本挑战赛不比拼纯算力或纯代码指标,而强调在复杂、高压的全球政治经济现实约束下,实现「社科战略判断力」与「AI 工具能力」的有机结合,切实打通出海企业「数据—洞察—决策」的实战链条。

关键信息

  • 主办单位北京大学国际关系学院
  • 承办单位全球风险政治分析实验室
  • 参赛对象全国高校本 / 硕 / 博研究生
  • 组队方式鼓励跨校、跨地区、跨专业组队
  • 赛事周期约 4 个月(共 15 周)
  • 报名细则官方通道将于近期公布

三大亮点

真题真做 · 人机协同 · 引才直通
01

重塑人机协同竞争力

文商科学生用 AI 放大战略直觉与逻辑画布;理工科学生跳出算法黑盒,在真实宏观博弈中淬炼算法的社会科学落地价值。

02

产业一线真题真做

告别离地学术假想,直接拿到中国全球化企业出海的真实痛点与脱敏数据,在极限高压下完成训练。

03

大厂与智库引才直通

优秀团队档案录入实验室长线追踪的「拔尖人才库」;决赛现场特设企业 1 对 1 专属引才室。

大赛介绍

Introduction首页 / 大赛介绍

大赛定位与宗旨

本赛事致力于发掘大模型及多智能体系统(MAS)在复杂网络中的认知增强与二阶冲击推演能力,推动「社科 + AI」的科研范式创新与拔尖人才培养,主动响应教育部「人工智能赋能高等教育转型」的号召。

01 — 办赛背景

认知增强 × 二阶冲击推演

将沉淀三十年的宏观地缘理论与国别经验,升华为护航中国企业全球化生存的「数智基础设施」,以赛代练,推动科研范式创新。

02 — 核心痛点

全球化的「隐性脆弱性」

深耕关键产业链节点的「专精特新」中小企业,因技术或供应链高度依赖全球协作,极易踩中大国技术封锁、长臂管辖、海外法律突变等风控红线。

参赛主体与意义

本届挑战赛面向全国高校和研究所的本科、硕士、博士研究生全面开放,专业不限。支持且鼓励跨校、跨地区、跨专业自由组队,实现多学科知识互补;高度欢迎留学生以本土化国别视角参与推演。

意义 01

人机协同复合竞争力

学会用 AI 放大战略直觉,或跳出算法黑盒淬炼社会科学落地价值。

意义 02

真题真做淬炼

直接拿到企业出海真实痛点与脱敏数据,在极限高压下完成训练。

意义 03

引才直通车

优秀团队录入拔尖人才库,决赛现场对接企业绿色招聘通道。

赛道设置

Tracks首页 / 赛道设置

双轨互补 · 长线长效 · 中段突攻

2 常规赛道 + 1 特别挑战

从宏观与微观两个维度,为企业出海做极限压力测试。

TRACK A

全球地缘政治不确定性量化感知赛道

Perception Track

聚焦「宏观全局层面」

利用大模型与 NLP 技术,对海量、异构的海外涉外法制动态与冲突文本进行全自动清洗与风险度量,转化为可量化的「地缘风险指数」。

TRACK B

专精特新全球化战略决策赛道

Strategic Decision Track

聚焦「微观行为体层面」

在「沙盒企业数据集(四大剧本)」约束下,利用 AI 模拟多智能体对抗博弈,推演不同生存路径的「风险-收益代价账单」。分低代码「业务咨询流」与高代码「硬核技术流」。

⬡ 低代码 · 业务咨询流⬡ 高代码 · 硬核技术流
72

地缘政治危机「人机交叉」推演赛道

Crisis Simulation · 主赛道中段无预警发布

组委会无预警注入海外高烈度地缘政治冲突,赛队须在 72 小时内迅速设计并搭建一套大模型多智能体系统(MAS)进行数字模拟,全自动推演多轮局势发展,得出熔断或撤离等应急决策。同时,主办方组织由政企专家、智库学者组成的「导演组真人模拟团队」进行实战对照,开展人机比较(但不以真人模拟为基准)。

【 72小时极限危机突发情境注入 】【 AI多智能体系统模拟 】参赛赛队基于真实推演流程搭建【 导演组特种真人模拟团队 】基于海保推演机制同步实战【 "人机协同VS真人推演"硬核交叉评估 】

赛程安排

Schedule首页 / 赛程安排

五阶段时间表 · 共 15 周

整体周期约 4 个月
1–4

阶段一启动与科普

发布一号公告与沙盒蓝本,发放基础地缘数据集,举办线上轻量化先修公开课(Coze、Dify 等低代码平台),完成跨学科组队与报名筛选。

5–10

阶段二赛事进行中

选手展开人机协同交互与战略推演;主赛道中段无预警发布「72 小时突发危机人机对抗极限推演」挑战任务。

11–12

阶段三双轨硬核评审

自研「Geo-Risk Auto-Eval」系统执行排重、合规与「防幻觉、查洗稿、验逻辑」红线审查;评审专家团完成第一轮盲审。

13–14

阶段四决赛总答辩

入围团队线下/线上汇报演示,政企学界联合评审团现场打分,评选特等奖及专门奖项并举行颁奖典礼。

15

阶段五成果落地展出

面向社会与企业展出获奖作品与产业转化落地,全面启动优秀选手引才对接。

技术平权 · 统一发放实验室封装的「基础地缘数据集」,集中举办低代码先修公开课,不设编程壁垒,重点考察人类的因果逻辑而非原生编程;启用自研「Geo-Risk Auto-Eval」实现「AI 辅助评测 + 人类专家把关」的双轨评审。

企业合作

Partnership首页 / 企业合作

资源对等 · 学术独立

大赛建立「资源对等、学术独立」的合作边界,外部技术与资金支持不得与赛事评判结果发生任何对价关系。合作单位可带回「四大实质资产」。

出海防御战略白皮书

联合智库专家编纂《企业日常海外运营风险连续性防范建议书》,直接作为企业风控决策参考。

风险量化工作流

高质量量化指数、风险传导画布及智能体压力测试工作流定向封装,协助企业搭建内部地缘风控系统原型。

数智复合型人才库

精准解决企业数字化转型与国际化进程中「既懂海外业务、又懂 AI 大模型」的人才荒。

国家级典型样板

脱敏前提下,优秀共创案例有机会编入北大新文科教研案例库,呈报国家相关决策部门。

合作单位权益

战略协办单位
Strategic Co-organizer

定向定制命题权、全国前三名 AI 工作流优先落地权、最高级别荣誉联合冠名与顶格宣发。

场景与案例提供方
Case Provider

专属出海风控工作流解决方案、优先调阅人才库、北大智库与技术专家联合拆解会诊。

技术与平台支持方
Tech / Platform Partner

细分领域「排他性指定技术 / 平台支持机构」荣誉冠名与全网联合宣发、高频深度品牌曝光。

填写企业合作意向 →

在线报名

Registration首页 / 学生团队报名

学生团队报名

2026 赛季 · 全国高校开放报名
队伍基本信息
团队 AI 工具经验 可多选
队长信息
队员信息 可继续添加,无上限
指导教师 选填,可多位

个人中心

My Account首页 / 个人中心

账号信息

参赛报名

完成账号设置后,即可提交参赛报名信息。

前往报名 →

企业合作联系

Partnership Inquiry首页 / 企业合作联系

填写合作意向

我们将在 2 个工作日内专人回复

欢迎企业 / 机构与全球风险政治分析实验室探讨战略合作,共同推动「社科 + AI」的产学研融合。

组委会合作联络 · 全球风险政治分析实验室
邮箱 riska@pku.edu.cn · 官网 riskalab.pku.edu.cn
30th Anniversary Special Event

Strategic Intelligence」Challenge

AI-Powered Geopolitical Risk Analysis & SME Globalization Decision-Making

Data · Insight · Decision

Days to Registration Deadline

Countdown

00D
00H
00M
00S
~4 months · 15 weeksB.S. / M.S. / Ph.D. All Welcome

About the Competition

At the historic convergence of the generative AI revolution and the global geopolitical transformation, Peking University's School of International Studies marks its 30th anniversary in 2026. Drawing on three decades of macrogeopolitical theory and country-specific expertise, the School is building a digital-intelligence infrastructure to guide Chinese companies navigating the global landscape.

This challenge is not about raw computing power or code metrics. It emphasizes combining social-science strategic judgment with AI tool capabilities under complex, high-pressure geopolitical and economic constraints — bridging the full Data→Insight→Decision chain for globally operating enterprises.

Key Facts

  • OrganizerSchool of International Studies, PKU
  • Co-organizerPKU Analytics Lab for Global Risk Politics
  • EligibilityAll university-level students nationwide
  • Team FormatCross-institution and interdisciplinary encouraged
  • Duration~4 months (15 weeks)
  • RegistrationDetails to be announced

Three Highlights

Real Problems · Human–AI Collaboration · Talent Fast-Track
01

Redefine Human–AI Synergy

Humanities students amplify strategic intuition with AI; STEM students move beyond black-box algorithms into real geopolitical scenarios that validate the social value of their work.

02

Real Industry Problems

Work directly with anonymized data and real pain points from Chinese enterprises expanding globally — under extreme time pressure, just like the real world.

03

Talent Fast-Track

Top teams enter the Lab's long-term talent pool; a dedicated 1-on-1 corporate recruitment session is held during the finals.

About

IntroductionHome / About

Competition Positioning

This competition seeks to uncover the cognitive-enhancement and second-order shock simulation capabilities of large language models and multi-agent systems (MAS) in complex networks. It advances the research paradigm of Social Science + AI and fosters top-tier interdisciplinary talent, responding to the Ministry of Education's call to leverage AI in transforming higher education.

01 — Background

Why Now

The simultaneous rise of generative AI and accelerating geopolitical fragmentation demands a new generation of professionals who can bridge the gap between strategic insight and AI-driven analysis.

02 — Methodology

Social Science × AI

Unlike pure hackathons, this competition requires teams to apply AI tools to real, declassified geopolitical data — evaluated on the quality of strategic reasoning, not just technical output.

03 — Vision

Bridges & Talent

Cultivate bilingual talent that speaks both the language of international relations and AI, and build institutional bridges between academia and globally operating enterprises.

Three Highlights

Real-World Stakes

Teams tackle actual de-identified business challenges from China's most globally active enterprises — not hypothetical case studies.

🤖

Human × Machine

The scoring rubric rewards strategic judgment and AI-augmented reasoning, explicitly valuing the interplay between human insight and tool capability.

🎯

Career Fast-Track

Outstanding teams gain access to the Lab's long-term talent tracking pool and direct corporate recruitment during the finals.

Tracks

Competition TracksHome / Tracks

Complementary Tracks · Long-Run & Sprint

2 Main Tracks + 1 Special Challenge

Extreme stress-testing of enterprise globalization from both macro and micro perspectives.

TRACK A

Geopolitical Uncertainty Quantification

Perception Track

Macro Global Perspective

Using large models and NLP, automatically clean and risk-score vast heterogeneous overseas legal and conflict texts, transforming them into a quantifiable Geopolitical Risk Index.

TRACK B

SME Globalization Risk Simulation

Decision Track

Micro Operational Perspective

Construct dynamic scenario trees for specific industries/markets, simulate multi-stakeholder strategic games, and output actionable risk control recommendations for outbound enterprises.

Special Challenge

Sudden-Crisis Sprint · Open to All
SPECIAL

72-Hour Crisis Simulation

Human vs. AI Extreme Deduction

Launched without warning mid-competition: an extreme 72-hour human-vs-AI crisis scenario. All teams participate. Results feed into the finals scoring as a wildcard.

Schedule

TimelineHome / Schedule

5-Phase Timeline · 15 Weeks

Total duration ~4 months
Weeks 1–4

Phase 1 Kickoff & Orientation

Release of Announcement No. 1 and the sandbox blueprint; distribute baseline geopolitical datasets; host lightweight online pre-courses on low-code AI platforms (Coze, Dify, etc.); complete cross-disciplinary team formation and registration screening.

Weeks 5–10

Phase 2 Active Competition

Teams conduct human–AI interaction and strategic simulations. Mid-competition, an unannounced 72-hour crisis human-vs-AI extreme deduction challenge is released.

Weeks 11–12

Phase 3 Rigorous Review

Proprietary Geo-Risk Auto-Eval system performs deduplication, compliance checks, and red-line audits (anti-hallucination, plagiarism detection, logic verification); expert panel completes the first round of blind review.

Weeks 13–14

Phase 4 Finals & Defense

Shortlisted teams present in person or online; a joint panel of academic, government, and industry experts scores in real time; Grand Prize and Special Awards presented at the closing ceremony.

Week 15

Phase 5 Results & Talent Matching

Public release of results; top teams enter the Lab's talent tracking pool; corporate 1-on-1 recruitment interviews; selected cases submitted for national policy reference.

Partnership

Enterprise CooperationHome / Partnership

Resource Parity · Academic Independence

The competition maintains strict boundaries: no external financial support may influence judging outcomes. Partner organizations gain four tangible assets in return.

Overseas Risk White Paper

Co-authored with think-tank experts: a Continuous Overseas Operational Risk Prevention Advisory for daily corporate risk management decisions.

Risk Quantification Workflow

High-quality risk indices, propagation canvases, and agent-based stress-test workflows — packaged to help build an internal geopolitical risk prototype.

Digital-Hybrid Talent Pool

Access to the Lab's curated talent database of professionals who combine global operations expertise with AI capabilities.

National Showcase Case

Outstanding co-created cases may be included in PKU's new liberal-arts research case library and submitted to national policy bodies.

Partnership Tiers

Strategic Co-organizer
战略协办单位

Exclusive rights to co-design competition themes; priority access to top-3 AI workflows; highest-tier co-branding and promotional partnership.

Case & Scenario Provider
场景与案例提供方

Bespoke overseas risk control workflow; priority access to talent pool; joint consultation with PKU think-tank and technical experts.

Tech & Platform Partner
技术与平台支持方

Exclusive sector designation as official technology/platform partner with full co-branding and broad promotional reach.

Submit Partnership Inquiry →

Registration

在线报名Home / Team Registration

Student Team Registration

2026 Season · Open to Universities Nationwide
Team Information
Team AI Experience Multiple selection
Team Leader
Team Members Add as many as needed
Faculty Advisor Optional, multiple allowed

Partnership Inquiry

Contact UsHome / Partnership Inquiry

Submit Your Inquiry

We will respond within 2 business days

We welcome enterprises and institutions to explore strategic partnerships with the PKU Analytics Lab for Global Risk Politics.

Contact · PKU Analytics Lab for Global Risk Politics
Email: riska@pku.edu.cn · riskalab.pku.edu.cn

My Account

ProfileHome / My Account

Account Information

首页 / 赛事新闻

赛事新闻

Competition News
2026-06-01
公告
「战略智能挑战赛」正式启动,面向全国高校开放报名
由北京大学国际关系学院主办、全球风险政治分析实验室承办的第一届「战略智能挑战赛」正式启动。大赛以 AI 赋能地缘政治风险分析为核心命题,设 Track A、Track B 及「72小时极限危机」特别赛道共三个方向,面向全国高校本科生及研究生招募参赛队伍。
Home / News

Competition News

赛事新闻
2026-06-01
NOTICE
Strategic Intelligence Challenge Officially Launched
The inaugural Strategic Intelligence Challenge, hosted by PKU School of International Studies and organized by the Analytics Lab for Global Risk Politics, is now open for registration. Three tracks: Track A, Track B, and the Special 72-Hour Crisis Track.
提交成功