'Flawed', 'Cruel' and 'Irresponsible': The Framing of Automated Decision-Making Technologies in the Australian Press

Abstract

This discussion paper outlines the findings of a qualitative news framing analysis of automated decision-making (ADM) technologies in the mainstream Australian press between 1997 and 2021. All articles including a reference to ADM in these media were identified using the Factiva news monitoring database. These articles, 40 in total, were then analysed for the ways ADM was framed, including paying attention to the headline, the broad and specific topics featured in each article, the news sources, whose interests received attention and the overall tenor of each article. For comparison purposes, I also conducted separate searches for each of the terms ‘artificial intelligence’, ‘algorithm/s, algorithmic’, robot/s/robotic’ and ‘face/facial recognition’. This analysis found a very low level of Australian news outlet reporting using the term ‘automated decision-making’. In some of the news stories, ADM was only briefly mentioned as part of broader discussions about software and big data. By contrast, other related digital technologies have been far more highly reported. The analysis also revealed that two news outlets (The Australian Financial Review and The Australian) and topics have dominated the very small corpus of Australian print news reporting specifically referring to ADM and that therefore, the vast majority of Australian news consumers would not have been exposed to such reports. The broad topics of services were most prevalent in the news reports. For the most part, rather than speculations about the futures of ADM employing techno-utopian or dystopian imaginaries, mundane services offered in the banking, financial, business or legal sectors offered by actually existing or near-future ADM received most attention. These ADM technologies were sometimes promoted positively in terms of benefits such as efficiencies and cost savings they could offer. However, they were far more frequently framed negatively in terms of actual or potential failures, mistakes, scandals or personal data privacy and security harms. Overall, ADM was positioned as untrustworthy and inferior to human decision-making, requiring close oversight by humans to ensure that Australians would not be disadvantaged or exploited by its irresponsible or thoughtless deployment by government or industry.