Search

From AI Capability to Stakeholder Confidence: The Trust Pillar of STEWARD-AI | nasscom | The Official Community of Indian IT Industry

Boards should embed AI trust from the start. The STEWARD-AI Trust pillar covers Transparency, Stakeholder Confidence, Reliability, and Explainability to ensure stakeholders can trust AI-enabled decisions through governance, communication, and ongoing risk monitoring.

The Trust Board: which AI tells you the truth?

The document presents benchmarking data on AI models' ability to answer questions truthfully, highlighting accuracy, hedged responses, and scoring limits, with mentions of GOV.UK as a data source and a governance-focused framing around trust in AI outputs.

Microsoft: 'Frontier AI models cannot be trusted

Translated title. Original: Microsoft: ‘Frontier AI-modellen zijn niet te vertrouwen’

Microsoft warns frontier AI models can be trusted only with safeguards; incidents show their power and potential misuse. They propose monitoring, protective agents, and real-time transparency, including reasoning processes, to ensure alignment with organizational goals.

Microcredentials Help Engineers Build Trustworthy AI - IEEE Spectrum

The article presents microcredentials as a way for working engineers to gain and verify skills in AI ethics, auditing, and compliance. It describes IEEE’s credentialing program as a route for applying standards such as the IEEE 7000 series to AI development.

[2610.30445] A Three-Pillar Governance Framework for Safe and Trustworthy AI Agents

A framework for safe, trustworthy AI agents uses a Three-Pillar model—transparency, accountability, and trustworthiness—to guide progressive autonomy with verifiable oversight, traceability, and auditable controls in enterprise settings.

How to trust AI agents without losing control | TI INSIDE Online

Translated title. Original: Como confiar em agentes de IA sem perder o controle | TI INSIDE Online

AI agents move from passive tools to autonomous systems, boosting productivity but requiring strong, early governance and security. Trust hinges on clear data access, deterministic controls, real-world testing, and maintaining human accountability with proportional supervision and evidence of safe operation.

Secure AI for Business: Keep Your Systems Safe

ZeroTrusted.ai provides a security layer that protects AI deployments by inspecting all interactions in real time, blocking prompts and jailbreak attempts, redacting sensitive data, and auditing actions. It works with existing models and deployments, requires no code changes, and offers compliance-ready testing and evidence generation.

Transparency in AI is crucial to the technology’s success • Table.Briefings

Translated title. Original: Transparenz bei KI ist entscheidend für den Erfolg der Technologie • Table.Briefings

Many companies promise human oversight of AI decisions but give little information about how people can challenge them. An AI expert warns that this gap can weaken trust and limit AI’s successful use.

DEBATE: Reply: To trust AI, we must also understand it

Translated title. Original: DEBATT: Replik: För att lita på AI måste vi också förstå den

The debate emphasizes trust in AI depends on safe data handling, clear governance, and staff AI competence; organizations must provide rules, oversight, and human review where needed, while ensuring users understand limits and risks.

How organizations progress from trusted data to trusted AI and agents | dbt Labs

The article argues that trusted AI and agents depend on reliable data, shared business context, clear ownership, and governance that scales with autonomy. It also says these capabilities must carry across systems through interoperable infrastructure.