iSAFE Hackathon
A global, hybrid hackathon to build trustworthy generative AI that defends truth, protects citizens, and promotes online peace.
An Official Pre-Summit Event of the AI Impact Summit 2026
36
Hours
20
Max Teams
5
Tracks
60
Participants
Why This Matters?
Generative AI can create content at scale — images, video, audio, and text. That capability is a force for good, and a source of new harms: deepfakes, targeted scams, automated disinformation, and model misuse. HackForTrust invites the global community to build practical, deployable tools that protect truth, safeguard citizens, and make AI accountable. This is a hands-on competition: prototypes must be demonstrable and ready for pilot testing.
Event Timeline
Your journey to building digital trust
November 20, 2025
Launch
Official launch of the Global iSAFE Hackathon
November 20, 2025 – January 15, 2026
Registration Period
Participants can register individually or as teams; open to global participants
January 15–25, 2026
Shortlisting Phase
Evaluation of submitted ideas/concepts; top entries are shortlisted for the finale
January 27, 2026
Finalists Announcement
Official announcement of teams selected for the Grand Finale
February 8–9, 2026
Grand Finale
36 Hours continuous development with mentorship and guidance from jury at New Delhi
February 10, 2026
Showcase @ Bharat Mandapam
Presentation of final solutions at Bharat Mandapam during the CyberPeace Summit, followed by recognition and awards
Competition Tracks
Choose your challenge and innovate
Detect the Deceptive
Tracing Synthetic Realities
Problem Statement
Synthetic content — AI-generated images, deepfakes, cloned voices, and fabricated texts — has blurred the line between truth and deception. The challenge is to design GenAI-powered tools that can detect, authenticate, and trace the origins of such content across platforms.
Core Challenge
Build systems capable of verifying digital media authenticity, providing content provenance, and flagging manipulative or synthetic content in real time.
Expected Outcome
A robust AI-driven verification system capable of identifying and classifying deceptive content with high accuracy.
Deliverables
- AI Watermark & Provenance Detectors: Verify content authenticity and trace its digital lineage
- Cross-Platform Deepfake Verifiers: Unified detection API for image, video, and text manipulations
- AI Truth Lens Browser Plug-ins: Real-time deception detection integrated with browsers and messengers
- Voice Clone Authenticity Checkers: Distinguish genuine human voices from AI-generated imitations
Defend the Digital Citizen
AI for Protection and Empowerment
Problem Statement
With GenAI accelerating scams, frauds, and misinformation, citizens need AI allies to stay informed and safe. The challenge invites solutions that use Generative AI to proactively educate, guide, and protect users from emerging digital threats.
Core Challenge
Design AI-powered educational or assistive systems that simulate real-world scam scenarios, train users to identify manipulative patterns, and offer personalized protection advice.
Expected Outcome
An AI-powered digital safety ecosystem that helps users recognize and avoid scams, enhances cyber awareness.
Deliverables
- Conversational AI Assistants for cyber hygiene and scam prevention
- Generative Role-Play Simulators to train users in recognizing phishing or AI manipulation
- AI Mentors for Vulnerable Groups — children, women, seniors — guiding them toward safe digital practices
Building the Guardrails of Trustworthy AI
Practical Red-Teaming & Governance
Problem Statement
As Large Language Models (LLMs) and Generative AI tools become more integrated into daily life, they also become potential vectors for misinformation, bias, and misuse. The challenge is to create real, testable safety guardrails that make AI systems safer, transparent, and aligned with human and societal values.
Core Challenge
Develop tools or frameworks that can test, monitor, and govern the behavior of AI models in real-world settings — making them more accountable and predictable.
Expected Outcome
A functional prototype or dashboard demonstrating how AI systems can be tested or audited for safety and alignment.
Deliverables
- AI Red Team Simulator: A sandbox where developers can test LLMs against prompt injection, jailbreaks, or harmful content scenarios
- Trustworthiness Scorecard: A dashboard that measures model safety, bias, and compliance based on inputs and outputs
- Ethics Layer Plugin: A middleware tool that sits between user prompts and model responses, filtering unsafe or misleading outputs
- LLM Audit API: API that logs and explains decisions made by the model to support transparency and regulatory auditing
Create for Peace
Generative AI for Social Good and Digital Civility
Problem Statement
Amid growing digital toxicity, misinformation, and polarization, this challenge reimagines Generative AI as a force for empathy, education, and peacebuilding. The goal is to create AI systems that inspire positive behavior and social resilience.
Core Challenge
Harness the creative power of GenAI to educate, unite, and heal — through storytelling, art, and immersive experiences promoting peace and digital ethics.
Expected Outcome
A GenAI-powered peace tech solution that inspires constructive dialogue and strengthens social harmony.
Deliverables
- Generative Storytelling Engines that teach empathy and digital harmony
- AI Campaigns for Digital Civility — narrative-driven outreach countering hate and misinformation
- Generative AR/VR Experiences for peace education and AI ethics learning
Enhance the CyberPeace Chatbot
Generative & Agentic AI Capabilities
Problem Statement
CyberPeace currently hosts a chatbot designed to promote online safety and responsible digital behavior. The next phase of innovation focuses on transforming this chatbot into an intelligent, context-aware, and proactive digital peace agent through the integration of Generative AI, CyberPeace GPT, and Agentic AI frameworks.
Core Challenge
Design and develop advanced modules that empower the CyberPeace Chatbot to provide real-time, adaptive, and personalized guidance to users on cyber safety, misinformation, mental health, and ethical AI use.
Expected Outcome
A working prototype of the upgraded CyberPeace Chatbot with integrated GenAI + Agentic AI capabilities.
Deliverables
- A working prototype of the upgraded CyberPeace Chatbot with integrated GenAI + Agentic AI
- Technical documentation explaining data pipelines, agent design, and ethical safeguards
- Dashboard for monitoring chatbot activity, safety analytics, and feedback loops
Registration & Submission
What you need to prepare
Team Formation
- Individual or team participation
- Maximum team size: 3 members
- Team-based participation encouraged
- 20 teams maximum for finale
Submission Requirements
- Abstract in PPT or PDF format
- Problem statement being addressed
- How the solution addresses the problem
- Concept and target users
Additional Submission Details
Evaluation Framework
How submissions are evaluated
Innovation & Technical Depth
Trust & Safety Integration
Societal Impact & Usability
Scalability & Localization
Presentation, Documentation & Ethics
The Grand Finale
During the Grand Finale, shortlisted participants will have 36 hours to transform their proposed solutions into working prototypes.
Day 1 - February 8, 2026
Jury meets teams, assesses initial ideas, provides mentorship and guidance for concept refinement
Day 2 - February 9, 2026
Continued development under mentorship, final presentations, Q&A sessions, and jury evaluation
Presentation Format
- 15 minutes total per team (10 min demo + 5 min Q&A)
- PowerPoint presentation (10-12 slides maximum)
- Live prototype demonstration required
- Evaluated on performance, speed, precision, and scalability
Format: Hybrid — Global online build + Finale at New Delhi + Opportunity to Showcase at Bharat Mandapam
Who Can Participate?
Open to innovators across nations
University Students
AI Researchers and Students
CyberPeace Corps
Volunteers and Innovators
Startups
Trust & Safety, Fintech, Media Tech
Defence & Police
Academics for Red-Team-Blue-Team Drills
Proposed Themes
Five powerful tracks to choose from
Detect the Deceptive
Trace, verify, and neutralize synthetic media (images, video, voice, text).
Defend the Digital Citizen
Build proactive GenAI tools that prevent scams, educate vulnerable users, and rebuild trust.
Building the Guardrails of Trustworthy AI
Practical red-teaming, audit, and real-time governance tools for LLMs.
Create for Peace
Generative experiences that teach empathy, counter polarization, and rebuild civic resilience.
Enhance the CyberPeace Chatbot
Integrate Generative AI, CyberPeace GPT, and Agentic AI to make CyberPeace's chatbot an active digital peace agent.
Prizes & Support
Cash Prizes
Winners will receive cash prizes and recognition
Cloud Credits
Selected teams can apply for partner cloud credits
Incubation Support
Fast-track incubation opportunities for winning teams
Pilot Projects
Opportunities for pilot projects with CyberPeace