Algorithmic Personalisation and Peace Journalism

The final speaker in this IAMCR 2019 session is Mariella Bastian, who points out the impact of the digital turn in journalistic conflict coverage; journalists themselves are now more mobile, but citizen content has also become easier to incorporate into the coverage. Further, digital media also intensify the dissemination of content coverage, and this could both increase or decrease hostility between the conflict parties.

Polarisation may also matter here, as it may aggravate the polarising effects of conflict-related news, open avenues for the manipulation of what content different users encounter, and undermine user trust in the quality of the content selection served to them.

One interesting approach in this context is peace journalism, which focusses on promoting peace and advancing conflict resolution; this means that peace (not war), truth (not propaganda), citizens (not elites), and resolutions (not victory) are central themes to such coverage.

Algorithmic personalisation may help or hinder in the attainment of these goals. It may help in recognising the complex relationships between the actors and factors behind these conflicts, but could also manipulate such understandings through the decontextualisation of reporting. Similarly, it may assist in increasing the diversity of voices heard about a given conflict, or lead to the formation of isolated ideological communities who are not exposed to oppositional voices. Further, it could enhance objectivity by promoting fact-based and balanced views of a conflict, or reinforce biases through selective exposure and a promotion of propaganda content. Finally, it might increase transparency by increasing audience trust towards constructive news stories if algorithmic functionality is transparent, or diminish the audiences sense of autonomy it it is not.

There are some positive opportunities for news personalisation in a peace journalism frame, therefore – but these opportunities can only be realised if the personalisation systems are appropriately designed and deployed. The importance of the integration or non-integration of normative values into algorithm design must be recognised here.