- Microsoft CEO Satya Nadella urges stronger safeguards and accountability in AI systems for businesses
- Nadella highlights the need to trace AI behaviour and disclose failures promptly in workplace AI
- He emphasizes building AI systems with controllable actions and transparent operations
Microsoft CEO Satya Nadella has called for stronger safeguards and greater accountability in artificial intelligence systems, arguing that businesses must be able to contain their actions, trace their behaviour and disclose failures in a timely manner as AI agents become more deeply embedded in workplace operations.
In a blog post discussing the evolving architecture of AI systems, Nadella stressed the need to reassess how trust is built into the next generation of technology. He called for mechanisms that allow developers and businesses to understand what went wrong when systems fail, alongside tools that can trace AI behaviour to specific code paths.
“We must build systems whose actions we can contain,” Nadella said, underlining the importance of control and transparency as organisations increasingly rely on AI to perform complex tasks.
ALSO READ: Microsoft's Decision-1 AI Model Claims 35x Faster Speed Than GPT-6 Sol, 2.5x Than H2O-Lightning-4B
He also highlighted the need for timely disclosures when AI systems malfunction and mechanisms to share information about failures. The emphasis on traceability comes as companies move beyond conventional AI assistants towards autonomous agents capable of executing multi-step workflows with limited human intervention.
Nadella's comments come amid Microsoft's broader push to reshape enterprise software around AI agents, with the company seeking to make Copilot a central interface for workplace tasks.
In an earlier discussion on the changing role of enterprise software, Nadella argued that AI could help people organise software around their work rather than forcing them to adapt to fragmented applications.
“So much of what we do at work is still spent working around software. We switch between apps, rebuild context each time, and translate what we want to do into the steps each system expects. AI is changing that,” he wrote.
Nadella pointed to activity on GitHub, Microsoft's software development platform, as an early indication of the shift. He said the growth of agentic development had coincided with accelerating repository creation and increased pull-request and commit activity.
According to Nadella, the rise of AI agents does not make traditional systems of record less relevant. Instead, it increases the importance of trusted platforms that maintain information, coordinate changes and manage business data.
Microsoft's strategy centres on combining Copilot's AI capabilities with business applications, organisational data and governed access to enterprise workflows. The company has also introduced tools designed to allow AI-generated code to run within corporate environments under IT oversight.
Nadella believes this integration could enable businesses to customise existing applications and develop new software experiences without fragmenting their core systems.
However, as AI systems gain greater access to sensitive business information and operational processes, ensuring that their actions remain traceable, controlled and accountable is becoming an increasingly important challenge.
His latest remarks underscore the need to build trust into AI infrastructure alongside expanding its capabilities, particularly as autonomous agents take on a growing role in enterprise decision-making and execution.
Essential Business Intelligence, Sharp Market Insights, Practical Personal Finance Advice, Daily Fuel, Gold and Silver Prices and Latest Stories — On NDTV Profit.