Infosys Team Wins Trustworthy AI Governance Hackathon
The win highlights new methods for controlling autonomous agents using the TM Forum AI-Native Open Digital Architecture Canvas.
Updated on Oct. 15, 2026 in Artificial Intelligence
Live Poll
Do you trust that businesses can implement effective safety controls for autonomous artificial intelligence?
Infyvibes, a team from Infosys, won the TM Forum Trustworthy AI & Data Hackathon held in Dallas. The event challenged 40 teams to build and verify controlled, agentic AI operations within an autonomous enterprise sandbox.
Why it matters
This event highlights a shift toward governing autonomous agent behavior through technical frameworks like the Open Digital Architecture Canvas. It aims to standardize how enterprises verify and control AI decision-making before deployment.
Participants executed over 35,000 governed model calls across 50 active ODA Canvas instances. The hackathon resulted in the creation of 196 custom AI agents, testing the limits of real-time enforcement and human-in-the-loop escalation systems.
The players
Infyvibes
An award-winning development team composed of engineers from the multinational IT consulting firm Infosys.
TM Forum
An industry association that develops standards for open digital architectures and software-defined networks.
AT&T
A global telecommunications provider that hosted the hackathon within its enterprise research environment.
The details
Developers worked within an autonomous enterprise sandbox to configure the TM Forum AI-Native Open Digital Architecture (ODA) Canvas, a framework for deploying modular software components. Teams connected various agents and tools to this architecture, implementing real-time enforcement protocols that monitor AI actions. To ensure safety, they integrated human-in-the-loop escalation—a system requiring manual verification before autonomous agents commit to critical operational decisions.
Timeline
The TM Forum Trustworthy AI & Data Hackathon occurred in Dallas on October 6-7, 2026.
The next TM Forum hackathon is scheduled for February 1-4, 2027, in Amsterdam.
The Tech Race
This event serves as a practical stress test for the TM Forum AI-Native Open Digital Architecture Canvas. It places the industry in a race to replace experimental AI prototypes with standardized, governable agentic workflows.
Enterprises currently using autonomous agents can look to these hackathon results as a blueprint for implementing human-in-the-loop controls. These validation methods are expected to become standard requirements for companies integrating AI agents into customer-facing operations.
The takeaway
The event proves that agentic decisions can be successfully governed through real-time enforcement and human verification. Industry practitioners should track the upcoming February 2027 event to see if these hackathon protocols are codified into official industry standards.
What happens next
The TM Forum will hold its next hackathon in Amsterdam from February 1-4, 2027.
Further reading
For more on how frameworks are evolving to handle automated decision-making, see our Artificial Intelligence coverage.
Source note: This article includes information reported by TelecomDrive.
Live Poll
Do you trust that businesses can implement effective safety controls for autonomous artificial intelligence?
