UNBIASED: Fact-provisioning and bias estimation tools for public inoculation against disinformation campaigns

Start Date: 01/11/2020
End Date: 30/04/2023
Project Leader: Papageorgiou Harris

The proposed project works in this direction by providing automated web tools for estimating the reporting bias and the polarization of news, or –more generally –information sources, as an indication of the “media bias”. To our knowledge, ICT projects have so far not approached this topic, partly because of the difficulty of estimating opinion bias, but also because correcting media bias is often regarded as something that is unattainable, or utopic (i.e. holding the view that media bias always existed, and will continue to exist). However, the abundance of web information and the capabilities of big data analysis have made it easier to detect different opinions and devise metrics for bias esti-mation.The project also aims at increasing automatization of fact-checking on debated topics. People make claims about facts all the time, but there is an unprecedented amount of half-truths, hyperboles, or straight-out lies. Fact-checkers today are mostly human-operated (e.g. web sites such as Politifact.com, Factcheck.org, fullfact.org), which generally results in an intellectually demanding, laborious and time-consuming process. There is, therefore, an increasing de-mand for automated fact checking, which can keep up with the rate at which false claims or hoaxes are produced on news sites and social media.The proposed tools are mainly targeted for the general public i.e. online news readers. The tools will enable the public to have more versatile and objective information, and to be aware of media bias in the news sources they choose for the daily information. Inaddition, it can help non-profit organizations that strive for more objective news coverage, (e.g. https://fair.org/), have automated tools for watching over the media. They can also help journalists and news agencies by enriching tools for collection, classification and analysis of news content, with new solutions for improv-ing the diversity and quality of news coverage; showing them which news topics have not received enough attention, providing them with external datasets so they can enrich a news article, etc.