The next presenter in this ICA 2024 conference session is Brian So, whose interest is in how Bloomberg is using automated reporting to cover the financial results of Hong Kong-listed companies. Automated reporting has long been seen as supporting especially sports, financial, and weather reporting, since reporting there tends to follow very formalised patterns.
Financial news is numbers-intensive, requires utmost accuracy, and highly repetitive, with earnings reports and regular updates put out by companies. News organisations tend to claim that their use of automated tools is not meant to replace journalists altogether, but to expand the scope of formalised coverage while freeing up journalists to engage in more in-depth reporting and analysis.
Bloomberg has been one of the earliest adopters of such technology, and even developed its own BloombergGPT tool. Some reporting is now done by machine only, some in a combined mode between humans and machine, and some continues to be done by human journalists. Machine reporting now emerges minutes after the initial financial reporting by companies; this is further added to by humans; and finally human journalists may publish further stories. Not all companies are covered in this way.
Brian applied news value theory to these stories, to understand what news values are present here. Key values explored here were especially social significance and deviance – are these covered disproportionately much by human reporting, compared with stories that address other news values? Or do other characteristics of companies, industries, or other aspects affect coverage patterns?
The study focussed on listed companies on the Hong Kong Stock Exchange over four years; some 40% of stories were written by machine, 33% by machine and human, and the rest only by humans. 33% of listed companies are not covered by Bloomberg at all, even if machine reporting would be possible here – this is surprising (deviance does not seem to be embedded as a news value in the algorithm). Larger revenue and higher stock prices attract more human reporting; sudden changes do not; financial industry companies tend to be covered more by humans, showing some industry bias.
