The One About Knowing When AI Is Ready
How do we know when an AI system is actually ready to be deployed?
Getting an AI system to work is one thing. Knowing whether it will hold up when it leaves development and enters the real world is another.
What needs to be tested before deployment? What changes when an AI system has to operate within a real organisation? And how do technical, operational and regulatory considerations come together when deciding whether a system is ready?
For this edition of AI Wednesdays, we're joined by members of the Sweden–Singapore AI Outlook Days 2026 delegation to explore that question from three perspectives: moving AI towards operational use in the public sector, building dependable AI systems, and testing AI through regulatory sandboxes.
More About The Sharings
Jon Karlfeldt (Programme Director, Swedish Public AI Works) will share on “From Mandate to Operational Use: Readiness in the Public Sector.”
What does it take to move AI from a government mandate towards something that can actually be used in practice? Jon will share how the Swedish Public AI Works attempts to enable AI adoption on the national, regional and local level of Sweden’s decentralised public administration. He will give examples of the practical challenges to moving AI systems to operational use and how the Public AI Works acts to address them.
Explore what readiness looks like when AI has to move beyond policy goals and into the realities of public-sector implementation, and what needs to be worked through before an AI system can be put into actual, real-world use. (Technical Level: 100)
Prof. Fredrik Heintz (Professor of Computer Science, Linköping University) will share on “Dependable AI: What It Takes for AI Systems to Hold Up When Conditions Change.”
How do we know whether an AI system will remain dependable when the conditions around it change? Fredrik will explore the technical side of AI readiness through the lens of validation, trustworthiness and resilience.
Examine what it means to establish confidence in an AI system before deployment, and the technical questions that arise when systems need to perform reliably beyond the conditions in which they were developed and tested. (Technical Level: TBC)
Dr Olle Engdegård (Affiliate Researcher, KTH Royal Institute of Technology) and Hans Hedin (Strategic Foresight and EU AI Act Innovation Support, Swedish Post and Telecom Authority) will share on “Inside a Regulatory Sandbox: What Gets Tested Before Deployment.”
What does an AI system need to demonstrate before it is ready to move beyond experimentation? Olle and Hans will take us inside the regulatory sandbox to explore how AI systems can be tested before deployment and what developers need to show along the way.
Explore how technical and regulatory considerations come together during testing, and how regulatory sandboxes can provide a space to work through those questions before AI systems are deployed in the real world. (Technical Level: 100)
More About The Speakers
Jon Karlfeldt is one of two Programme Directors leading Swedish Public AI Works, a government-commissioned initiative led by the Swedish Social Insurance Agency and Swedish Tax Agency. The initiative is developing shared infrastructure, expert support and reusable services to help agencies, municipalities and regions build, test and use AI securely.
Prof. Fredrik Heintz is Professor of Computer Science at Linköping University, where he directs the AI4X Center of Excellence, the Division of Artificial Intelligence and Integrated Computer Systems, and the Reasoning and Learning Lab. His research focuses on artificial intelligence, particularly trustworthy AI, the intersection of machine reasoning and machine learning, and autonomous systems.
Dr Olle Engdegård is an Affiliate Researcher at KTH Royal Institute of Technology, where he works on AI regulatory sandboxes and innovation support. His work looks at how sandboxes can combine legal advice with technical and organisational support across the innovation lifecycle.
Hans Hedin works on strategic foresight, intelligence analysis and innovation support related to AI, digitalisation and electronic communications at the Swedish Post and Telecom Authority. His work includes supporting the implementation of the EU AI Act while fostering innovation and responsible AI adoption.
More About The Series
AI Wednesdays is Lorong AI’s weekly gathering, bringing together practitioners, researchers and innovators for technical discussions on research insights, product development and engineering practices.
Get involved: Learn more about Lorong AI | Speaker Sign-up | WhatsApp Community | LinkedIn | X
