Tweet a lie and you’ll get caught!
You will soon be able to tell if a tweet is telling the truth or lying, thanks to a new ‘lie-detector’ application that scientists in the UK and Nairobi are developing.
The application dubbed ‘Pheme’ will have the capacity to analyse 140 characters or less and tell if they are factual or just plain rumours ‘in real time’. This is set to change the digital age for good and hopes to bring the rumours, lies and misinformation on Twitter to a screeching halt.
The program is a collaboration of five UK-based universities and several technology companies including Nairobi’s i-Hub, which is based on Ngong Road.
According to i-Hub’s Nathaniel Manning, the firm is working in partnership with its sister organisation Ushahidi and the Pheme consortium with the support of the European Commission to deliver the project.
“We are focused specifically on using the lie-detecting ontologies and algorithms created by the consortium and plugging them into the SwiftRiver platform that Ushahidi has developed,” said Manning.
The program will pay special attention to digital journalism and more specifically, to support journalists to ease their verification processes through automatic methods for detecting rumours and misinformation in user-generated content (UGC), modeling authority in social networks, and tracking information diffusion across social and traditional media.
In an interview with UK’s The Times newspaper, the lead researcher Kalina Bontcheva from the University of Sheffield, says the application will not only be able to test the information, but also track its origin.
She added that this application will come in handy, particularly in critical situations such as political riots and elections.
“In critical situations, you can instead show reliable information or alert the authorities before things get out of hand,” she said.
The three-year project, which commenced this month will classify online rumours into four types; speculation, controversy, misinformation and disinformation.
The downside, however, could be that people online could take to using vernacular and sending coded messages in a bid to circumvent the watchdog.
“There is likelihood that we will have a lot of pretenders online and it will push people to use vernacular and coded messages.
That is a threat they need to consider,” said Alex Gakuru, the chairman of ICT Consumers Association of Kenya.
The other downside, he said, is that computers were not always able to contextualise certain words and phrases deemed as inflammatory or hate speech.
While monitoring will work to keep bullies and liars off the online streets, Mr Gakuru points out that there could be another spectrum to the program, which is surveillance.