In the Wednesday episode of Portfolio's Checklist podcast (the segment begins at 14:53), Kiss Gergely, a member of the board at Signifer Group who focuses on corporate digitalization and artificial intelligence, discussed what kinds of information large language models collect and how they infer personal details from users.
He noted that roughly one third of Hungary's population already uses such models, making the topic socially significant. These algorithms do not rely solely on the literal content of questions: they also draw signals from phrasing, style and recurring topics. From those cues, models can infer aspects such as socioeconomic status, private-life details and behavioral patterns.
Earlier in the episode the hosts addressed a separate agricultural debate—partly covered on the Portfolio Group's Agrárszektor agricultural portal—about which new crops might be suitable for the increasingly drought-prone Great Plain (Alföld) and how prepared the processing and food industries are to handle any shift in cultivation. Veres Virág Cintia, a journalist at Agrárszektor, was interviewed on that topic.
Where to listen and related information
The episode is available on Spotify, Apple Podcasts and other major podcast platforms, as well as via the embedded player in the article. The main timestamps in the Checklist episode are:
- Intro – (00:00)
- Which crops could suit the drying Great Plain – (01:50)
- How much AI knows about us – (14:53)
The article also mentions that Portfolio will hold an AI & Digital Transformation 2026 conference on November 26; registration and event details are available on Portfolio's event page.
Finally, Portfolio is asking readers to help improve its subscriber program by completing a questionnaire; respondents can enter a draw to win one of five pairs of conference tickets.
The discussion highlights a practical concern: as large language models become more widespread, the linguistic signals we emit can be used to build detailed profiles, raising data protection and ethical questions that merit public attention.
