Independent research foundation · Romania

Research with purpose. Innovation with impact.

We are an independent research foundation dedicated to advancing artificial intelligence and robotics — pursuing rigorous, original research with real-world relevance and lasting societal value.

Robotics and AI research laboratory

Our story

Born from a conviction that great research should change the world

The Robotics & AI Research Institute was established with a clear purpose: to advance the frontiers of artificial intelligence and robotics through rigorous, independent research — and to ensure that this research generates knowledge with genuine relevance beyond the laboratory.

Too often, breakthrough research remains confined to academic papers. We believe that the true measure of research excellence lies not only in the quality of its publications, but in the depth of its impact — in factories, hospitals, boardrooms, and in everyday human interactions with intelligent machines.

Based in Romania and operating as part of Fundația Institutul de Cercetare în Robotică și AI, we bring together researchers and domain specialists from across academia and industry to investigate questions like:

How can robots and humans interact and form trust?

How can AI support better decisions without replacing human judgement?

How do organisations avoid losing knowledge when people leave?

Our values

What guides every programme we undertake

Real-world relevance

Every research programme we undertake is evaluated not only for scientific merit, but for the tangible problems it can help address outside the lab.

Human-centred research

Intelligent systems should work with people, not around them. We investigate transparency, trust, and human oversight at every level.

Openness and collaboration

We actively seek partnerships with enterprises, universities, and public institutions — because the hardest problems rarely yield to a single perspective.

Responsible research

We take seriously the ethical dimensions of AI and robotics — questions of fairness, accountability, and societal impact are not afterthoughts, but integral to our work.

Accessible knowledge

We communicate our findings in plain language, making our research useful to practitioners, policymakers, and the general public — not only to specialists.

Research with purpose

Innovation with impact — the standard against which every programme at the institute is measured.

Research areas

Four frontiers of intelligent systems research

01 — Research Area

Human–Robot Interaction

We study how people and robots share space, communicate intent, and form working relationships — spanning natural language interfaces, gesture recognition, adaptive robot behaviour, and safe collaboration in real environments.

Illustration: human–robot interaction

02 — Research Area

AI Decision-Making for Enterprise

We research AI systems that synthesise data from across siloed enterprise environments — CRMs, ERPs, document stores — producing sourced, auditable answers and supporting decisions without replacing the humans who make them.

Illustration: AI decision-making for enterprise

03 — Research Area

Industrial Automation & AI-Driven Production

We investigate adaptive automation systems in collaboration with industrial partners — studying how AI-driven machines learn from operational experience and work naturally alongside human operators on the floor.

Illustration: industrial automation and AI-driven production

04 — Research Area

Organisational Knowledge AI: reliable, grounded, self-healing

Modern enterprises run on fragmented knowledge — spread across CRMs, ERPs, email archives, wikis, and the tacit expertise of individuals. When people leave, that knowledge walks out with them. When systems don't communicate, decisions are made on incomplete information.

We research AI architectures that unify this fragmented landscape, querying all connected systems simultaneously, synthesising grounded answers, and continuously auditing themselves for contradictions, gaps, and outdated content.

Self-auditing knowledge bases

AI that detects semantic duplicates, contradictions, and unanswered questions — flagging them for resolution rather than propagating errors.

Zero-hallucination architectures

Grounding mechanisms that prevent AI systems from generating plausible answers unsupported by verified organisational knowledge.

Cross-system natural language query

Interfaces that let non-technical users query all connected enterprise systems in plain language — no SQL, no filters, no training required.

Cited, auditable AI responses

Architectures where every AI-generated answer is accompanied by traceable sources — essential for trust, compliance, and review.

A core principle: AI that doesn't know something should say so — not fabricate a confident-sounding answer.