AI Must Not Outrun Oversight: Andonovski Calls for Responsible Development at Impact Bucharest

Responsible AI at Executive Speed: Lessons from Impact Summit Bucharest

At the Impact Summit in Bucharest, Minister for Digital Transformation Stefan Andonovski stressed a simple truth: AI can transform economies and societies, but its development must not outpace our ability to manage it responsibly. The panel asked a blunt question: "Artificial Intelligence - revolution, apocalypse or big disappointment?"

Andonovski underscored the government's push to build legal, ethical, and technological guardrails for safe and inclusive AI use in the public sector. For managers, this signals a near-term shift in expectations: clearer rules, higher standards, and stronger oversight for AI in operations, procurement, and reporting.

Michelle Obama also addressed the event, focusing on leadership, innovation, and how new technologies impact people. The message to leaders was implicit: technology moves fast; your culture, decisions, and processes need to keep pace without losing trust.

What this means for managers

• Set a company-wide AI policy now. Define allowed use cases, data sources, model access, and human oversight.
• Create an AI governance group with IT, legal, risk, HR, and line-of-business leaders. Give it decision rights and a clear escalation path.
• Adopt a risk framework for consistency. The NIST AI Risk Management Framework and the EU's approach to AI policy and compliance are practical starting points.
• Run small pilots with narrow scope and clear exit criteria. Log prompts, decisions, and results for auditability.
• Tighten data hygiene. Classify sensitive data, restrict training on proprietary datasets, and document lineage.
• Vet vendors beyond demos. Ask for model cards, security practices, evaluation results, and update cadences.
• Define KPIs that blend efficiency and risk: accuracy, bias checks, incident rate, review time, and user satisfaction.
• Upskill the workforce. Train managers, analysts, and creators on practical AI workflows and policy compliance. Explore role-based paths here: AI courses by job and popular certifications.

Why the public sector matters to you

Public sector adoption sets the tone for procurement standards, audit requirements, and transparency. As governments formalize expectations, private-sector contracts will reflect similar demands around documentation, risk controls, and explainability.

Plan for alignment now. Map your current and planned AI use cases to expected regulatory categories, required controls, and reporting obligations.

30-day action plan

• Inventory AI use across teams (tools, data, decisions affected).
• Publish a one-page AI policy and a simple approval workflow.
• Pick two use cases for pilots; define success metrics and risk checks.
• Stand up an AI risk register and log incidents or model changes.
• Schedule manager training and create a feedback loop for frontline teams.

The takeaway from Bucharest is clear: speed with guardrails. Leaders who pair ambition with governance will compound gains while protecting customers, teams, and brand trust.