Can You Prove Your AI Can Be Trusted?

Key takeaways 4 min read

Bring your team to Dallas, Oct. 4–7: CSP and telecom vendor developers, architects, AI practitioners and data specialists can register now for the TM Forum Innovate Americas Hackathon to build trusted AI alongside Dell Technologies.

Prove trust, not just capability: Autonomous AI succeeds when teams can show what an agent did, why and whether it holds up to scrutiny.

Follow the evidence with “Agentic Assurance”: The hackathon challenge rewards clear, verifiable proof that AI decisions align with governance and business intent.

Quick answer: The TM Forum Trustworthy AI and Data Hackathon, part of Innovate Americas 2026, challenges telecom professionals to prove AI trustworthiness through evidence, traceability and explainability, rather than simply building AI capabilities. It runs October 4–7 in Dallas, Texas with Dell Technologies as a named partner.

Trust is one of the defining challenges of autonomous AI — not capability, not speed — trust. As communications service providers (CSPs) deploy AI-native architectures across customer, network, IT and business operations, a critical question keeps surfacing: how do you prove that an autonomous agent did the right thing, for the right reason, with the right outcome? Asserting that AI is trustworthy isn’t enough — organizations need evidence.

That’s the premise behind the TM Forum Trustworthy AI and Data Hackathon, taking place October 4–7, 2026, in Dallas, Texas. Dell Technologies is proud to be a named partner of this challenge, alongside Accenture, AWS, NVIDIA and ServiceNow. I’m sharing this because the hackathon addresses something I care deeply about: building AI that telecoms can actually rely on.

What makes this hackathon different from others?

Most hackathons measure success by what AI can build. This one measures it by what teams can prove.

Participants enter a simulated telecom-oriented environment containing pre-built agents, APIs, ODA Components, shared platform services and a simulated customer-to-network operational workflow. The challenge, titled “Agentic Assurance: The Quest for Proof,” asks teams to reconstruct what autonomous agents did, why they did it and whether their actions can be independently verified.

Teams are not building new telecom functionality. They’re doing something harder: creating explainable chains of evidence across a functioning autonomous enterprise. Success is measured by the strength of that evidence, not the sophistication of the AI itself.

This distinction matters. In production environments, the ability to explain an AI-driven decision, trace it back to specific data and policies and verify it against governance intent is what separates deployable AI from experimental AI.

How does the hackathon challenge work?

The hackathon uses TM Forum’s AI-Native ODA principles and standards, applied through the “Breaking Barriers” framework. Teams work across three organizer-controlled alert levels that simulate progressively complex conditions:

AL0, Baseline: A functioning fault management scenario with no injected threat.

AL1, Degraded: The baseline scenario is subjected to reliability or decision-quality degradation.

AL2, Compromised: The scenario introduces deliberately malicious behavior, including potential audit-trail tampering.

At each level, teams must identify risks, apply guardrails, detect unreliable or manipulated behavior and reconstruct what happened using telemetry, system logs, integrity checks and operational signals. The winning team will be the one that provides the strongest, most independently verifiable evidence trail.

Who should participate in the trustworthy AI and data hackathon?

This challenge is designed for organizations with direct relevance to the telecom industry. CSPs, telecom vendors, system integrators, AI software vendors serving the telecom sector and start-ups shaping trusted AI innovation are all encouraged to apply.

Strong teams will bring a mix of expertise across:

AI and machine learning

Data science and analytics

Governance and compliance

APIs and integration

Observability and monitoring

Security and identity

Telecommunications and enterprise operations

Teams consist of three to seven members. Companies may enter multiple teams. Participation can be in-person, hybrid or virtual, making this accessible regardless of geography.

Key dates and locations

The hackathon kicks off Sunday, October 4, at AT&T Corporate Headquarters in Dallas, Texas with the support of AT&T. Teams then move to the Marriott Dallas Allen when Innovate Americas 2026 opens Tuesday, October 6. Winners are announced during the keynote on Wednesday, October 7, in front of the full event audience.

Spots are limited for both in-person and virtual participation. TM Forum reviews all submissions, and placement is confirmed only upon direct notification. Given the timeline, teams are encouraged to submit their interest soon.

Why this challenge reflects the future of AI in telecom

Agentic AI is no longer a future scenario. Autonomous agents are making decisions across customer operations, network management and IT environments right now. The question of governance, explainability and accountability isn’t theoretical. It’s operational.

At Dell Technologies, we believe that AI adoption at enterprise scale requires more than performance benchmarks; it requires proof. Proof that decisions are traceable. Proof that governance controls are working. Proof that when something goes wrong, organizations can reconstruct exactly what happened and why.

The Trustworthy AI and Data Hackathon creates a structured environment to develop and demonstrate exactly that capability. It’s one of the most relevant challenges facing our industry, and I think the teams who engage seriously with it will come away with skills and frameworks that apply directly to their work.