And the final speaker in this entertaining session at the IAMCR 2025 conference in Singapore is Junbin Su, whose interest is in algorithmic gatekeeping in AI news recommendations. What values guide such gatekeeping decisions?
Conventional journalistic gatekeeping reflects editorial decisions about news values, and a number of competing lists of such news values have been proposed over the decades. Algorithmic gatekeeping may build on these values, or introduce others – not least also linked to platform metrics like user engagement or shareworthiness. This might under- or overemphasise certain news values.
This study explored this by identifying some 100 news recommender patents from Chinese, EU, and US patent databases, and explored what news values they claimed to embed in their processes. It found that personalisation-oriented rather than content-driven news values played a strong role, and that they introduced processes for ranking and filtering news based on structured data rather than human judgments. This reduces the emphasis on the editorial quality of the content itself. It is important to note that patents only reflect the intended design, however, and not necessarily everyday use patterns.
