Data scientist Suad Aldarra is currently trying to understand and highlight the role traditional media is having in stoking fears about refugees.
The online world in which many of us spend a large part of our day has become increasingly hateful. Whether it’s people on Facebook sharing posts about their distrust of others from minority backgrounds, or an orchestrated effort to change the social fabric of society, there is an obvious problem.
Yet for Syrian data scientist Suad Aldarra, the role of the traditional media – the journals of record that many of us would think of as following a high standard – is not studied enough when it comes to the spreading of hate speech against refugees.
“The thing is most of the research [examining hate speech] is focused on social media, not on news media,” Aldarra said in conversation with Siliconrepublic.com.
“In social media, it’s easier to detect because the language is less formal. There’s a big resource for hate words that you can detect, but in the news it’s more difficult because it’s formal and structured sometimes in a hidden way.”
From Egypt to Galway
Aldarra was one of millions of Syrians to leave her country in 2012 – along with her husband – following the outbreak of the nation’s civil war. Not long after moving to Egypt, the military coup there in 2013 meant she and her husband could no longer get a work visa, requiring her to once again look for a new home.
“We had to find another option,” she said. “I was applying for jobs everywhere … and I eventually got accepted for a job as a research software engineer with Fujitsu and moved to Ireland in 2014.”
The move from Egypt to Galway was certainly a change of scenery, in many ways, according to Aldarra.
“I remember everything was green when I first moved,” she said. “The weather is still something I struggle with but, as a culture, people were really nice and accepting and they were really welcoming.”
While being a data scientist was not in her career plans, she said that she became “hooked” on the idea of using data for good after watching a talk at the Web Summit in Dublin involving UNICEF and an organisation called Techfugees.
The latter was founded in 2015 to find ways to use tech to aid the plight of refugees across the world. After conversations with UNICEF, Aldarra joined as a volunteer to put her data science skills to use, by analysing data from social media pages to try and map the movement of refugees through Europe.
Not long after, Aldarra joined UNICEF full time, but she is now currently leading the Ireland chapter of Techfugees. While at UNICEF, one of the data science projects Aldarra worked on was ‘Refugees Are’, a website that scoured traditional media using an open-source tool called GDELT. This is one of the biggest open-source collections of news articles in more than 100 languages.
Collating news over the course of a month, Aldarra searched for common words used in articles when describing refugees. Unfortunately, much of the news coverage she witnessed in her research was negative – and quite strongly in some cases.
“I think I reached a level where I wasn’t surprised anymore,” she said. “There was a lot of hate messages.”
‘Sometimes people don’t realise when you read a headline that it is xenophobic or it contains a hidden message’
– SUAD ALDARRA
Her subsequent attempts to train a classifier to detect hate messages in media are focused largely on the three most read news sources in UK, which have made claims such as refugees were stealing jobs from UK nationals and were using up taxpayers’ money.
“All those topics were in the headlines and unfortunately that’s what is spreading through the media,” Aldarra said.
“The three news sources I looked at were the most read publications in the UK. You can imagine the spread of negative messaging to people who, for example, have never met a refugee but that’s all they hear about every day.”
The hidden messages in headlines
In a blog post documenting her search through media, Aldarra explained why she thinks her research is important : “I am angry and frustrated and disappointed most of the time, more now that I live in Europe and see how unfair life can be for people born and stuck on the wrong side of the borders,” she wrote.
“But I naively keep trying to make a change. Make things right. Bridge the gap and save the world that sometimes looks like it doesn’t want to be saved.”
After setting up Refugees Are in 2018, Aldarra brought her project to the Techfugees Challenge. She was one of 25 finalists chosen to pitch their project at the Techfugees Summit that year, and was also crowned the winner of the awards’ social inclusion category.
Speaking on what she hopes to do now with her research, Aldarra said: “I want to make [my news classifier] as a tool that can be used by agencies – as well as the public – to check the news today and give them a different perspective, something to think of when they look at the news.
“Sometimes people don’t realise when you read a headline that it is xenophobic or it contains a hidden message. Using this tool would help them think more about the article. I also want to apply it to Irish news, which is not explored, especially since there seems to be an interest in legalising against hate speech in Ireland.”
Want stories like this and more direct to your inbox? Sign up for Tech Trends, Silicon Republic’s weekly digest of need-to-know tech news.