Professor Ehud Reiter, a long serving academic at the University of Aberdeen, is widely regarded as one of the pioneers of Natural Language Generation (NLG) - the technology that underpins today's most widely used AI systems.
As he prepares to retire, colleagues, students, and industry partners are reflecting on a career defined by a simple but transformative question: how can computers learn to write like people to generate real-world impact?
A career built on turning data into language
Professor Reiter joined the University of Aberdeen in the 1990s, becoming a central figure in the emerging field of computational linguistics. At a time when most AI research focused on recognising patterns or classifying information, he championed a different frontier - teaching machines to express information clearly, accurately, and naturally.
Some of his most influential contributions include:
- Rule‑based NLG systems - early frameworks that allowed computers to turn structured data into readable text.
- Bridging linguistics and AI - establishing NLG as a discipline in its own right – and breaking down boundaries between the humanities and Computing Science, working with students and collaborators on Latin, Arabic dialects, and Syriac, among other languages
- Advocating for rigour in system evaluation – emphasising the importance of real-world evaluations and meaningful scientific measurements
- Mentoring generations of researchers - many of today’s NLG specialists trained under him or were influenced by his publications.
Reiter literally wrote the book on NLG—twice! The first, published in 2000 with co-author Robert Dale, and the second, authored solo in 2025, are the definitive books on the subject. Reiter’s work was never just theoretical: the first book focused on practical aspects of building NLG systems and the latest book focuses on the high-level questions that remain the same whether you are using rule-based systems, modern LLMs, or the next revolutionary AI technology. He consistently applied NLG to real‑world problems, especially in healthcare and accessibility.
Research helping people with everyday life
Throughout his career, Professor Reiter led projects that used NLG to support groups who often struggle to access clear, personalised information.
- Helping children with complex communication needs - his team developed technology that automatically generated personal stories for children with disabilities, helping them communicate more confidently with teachers and peers.
- Improving communication for neonatal care - Reiter also contributed to research that generated personalised updates for parents with babies in neonatal units. These automatically written summaries helped families understand complex medical information at an incredibly stressful time, reducing anxiety and improving communication between clinicians and parents.
Modern AI models, including the conversational systems people interact with every day, rely heavily on principles that Professor Reiter and his peers helped define. Before large language models existed, his research laid the groundwork for:
- how machines decide what to say
- how they choose how to say it
- how they adapt tone, structure, and clarity for different audiences
These issues continue to shape NLG systems today as they become more widely adopted in the era of generative AI. Prof. Reiter’s emphasis on real-world evaluations now seems prescient as these technologies have greater real-world impacts.
Arria NLG: From University Lab to global impact
Reiter was also a co‑founder and Chief Scientist of Arria NLG, a University of Aberdeen spinout that became one of the world’s leading NLG companies.
Arria’s technology has been used to automatically generate:
- financial reports
- business intelligence summaries
- weather forecasts
The company helped bring NLG out of academia and into everyday commercial use, demonstrating the real‑world value of Professor Reiter’s ideas long before AI became mainstream.
A legacy written in every AI‑generated sentence
While the public often associates AI with the latest breakthroughs, the field is built on decades of patient, rigorous work by researchers like Professor Reiter. His influence is present every time someone asks an AI system a question and receives a clear, coherent answer.
His retirement marks the close of an extraordinary chapter, but his ideas, students, and technologies will continue shaping how the world communicates with machines for decades to come.
Celebration
Professor Reiter’s storied career and his many contributions to the research community will be celebrated as part of a two-day symposium on Natural Language Generation Evaluations.
The event at Sir Duncan Rice Library will include talks, posters, panel discussions, invited keynote speakers and an outing to Dunnottar Castle and a celebratory dinner.
More information is available here: https://retroeval.github.io/