Newsroom
Latest news from the alignAI doctoral network:
When AI Learns Something New, Does It Stay Safe? AI Safety at ACL 2026
The 64th Annual Meeting of the Association for Computational Linguistics (ACL 2026) took place in San Diego from 2–7 July 2026. As one of the leading international conferences in natural language processing, ACL brought together researchers from around the world to exchange ideas on large language models, AI safety, alignment, evaluation and other emerging directions in language technologies. alignAI Doctoral Candidate Yung-Chen Tang attended the conference and presented the paper “Why LLM Safety Guardrails Collapse After Fine-tuning: A Similarity Analysis Between Alignment and Fine-tuning Datasets”.
Whose Values Shape the Model? — On Situating AI Alignment
In the film Humans in the Loop, there is a scene with a woman named Nehma, who works as a data labeller in a rural data-labelling centre in India. Nehma is tasked with annotating images of insects so that an AI-powered agricultural machine can identify and target them. Her American client needs her to label these insects as pests. When shown an image of a leaf-eating caterpillar, Nehma refuses to mark it as a pest. Drawing on traditional ecological knowledge, she explains that the caterpillar only eats the decaying parts of the leaf, allowing the rest of the plant to survive and thrive. Her manager tells her to follow the guidelines strictly, rather than applying her own understanding of what counts as a pest.
Why We Bond with Chatbots: The Psychology Behind AI Mental Health Tools
In 1956, researchers Horton and Wohl described something they had noticed about television audiences: people were forming genuine emotional bonds with personalities they had never met and never would. They felt affection, familiarity, even concern for figures who had no awareness of their existence. Horton and Wohl called these parasocial relationships and at the time the phenomenon was mostly treated as a curiosity of broadcast media.
Self-aware AI…or Just Reasoning Machines?
When the Terminator films first depicted Skynet becoming self-aware, the moment was presented as a turning point not only for humanity, but for the idea of intelligence itself. In the film, this “self-awareness” meant more than simply processing information or adapting to input. It signified the recognition of one’s own existence, the ability to formulate independent goals and the emergence of a self-preservative instinct.
alignAI’s Third Seasonal School in Eindhoven Puts Explainability and Fairness at the Centre of LLM Design
From 15 to 19 June 2026, the alignAI Doctoral Network held its third Seasonal School in Eindhoven. The five-day program was organised by researchers and staff of Eindhoven University of Technology (TU/e), and co-hosted by TU/e and the Jheronimus Academy of Data Science (JADS). Sixteen alignAI doctoral candidates (DCs) took part, together with three principal investigators, additional researchers and staff members.
Who Reads the News? On Asking and Reading in the Age of Conversational AI
For most of the history of modern journalism, “reading the news” has been understood as an active, often solitary, one to one encounter between a reader and an author’s text. Even in its most passive forms, such as scrolling through a feed and browsing through headlines in the morning, reading implied a degree of direct contact with journalistic output, hence with the words a journalist chose, the framing an editor approved and the story a newsroom decided to publish.