aiah dance”, a phrase which would throw up the video of a farmer, bearing an uncanny resemblance to the state’s chief minister, dancing to a raunchy Kannada song.
The clip was widely circulated on social media in an attempt to discredit Siddaramaiah, who is looking to win a second term in the assembly elections due in May. The spread of this fake news was halted only after a farmer from the state’s Mysuru region owned up that it was him and not the CM doing a jig. The ‘Siddaramaiah dance’ is only the latest in a slew of false news that has been spreading like wildfire on social media.
India has the largest user base for social networks including Facebook, WhatsApp, its messaging and payment platform, YouTube, the Google-owned video platform as well as Twitter, the micro blogging platform. Millions use these platforms to either consume or share content with friends. This massive user base also serves as a prime channel for those looking to spread misinformation with impunity.
Experts fear that as the country heads towards a series of state elections and in 2019, the general elections, tools that allow for targeted news delivery to specific demographics can be misused by interested parties. “Governments should put pressure on these companies to make the information available on their systems more trustworthy,” said Soumitra Dutta, professor of operations, technology and information management at Cornell SC Johnson College of Business. Dutta is of the view that machine dominated attempts to screen news and information in online platforms will succeed only to a degree.
“Machine learning algorithms also have their biases and humans have to be aware of these limitations. Social media companies will need to make more investments in screening online information,” he said.
Social media companies claim they use a mix of technologies such as machine learning and artificial intelligence along with the human review system to curb false news being pushed, experts argue that it is not enough. In most case, the platforms are unable to flag or control such news and when they do the process take up a lot of time before the action is finally taken, they say.
Jency Jacob, MD of Boom, a fact-checking initiative said that to a large extent companies are trying to do deal with the issue algorithmically. “But people have learnt to hoodwink the system, if you are part of the ideology you can create a debate and narration around it without getting flagged. There will be no abuses, no slang so obviously the automation is not able to catch them,” he said.
Facebook, which has been under scrutiny over the data of its users being wrongly used to influence elections says it will tighten controls, including sending forensic auditors to analyse user data that has been used by app developers. Mark Zuckerberg, its founder has apologised for the damage that it has caused in influencing elections. The company is working hard to improve the accuracy of information on Facebook, a spokesperson told ET. “We use various signals to predict which stories may be false or misleading.
One of the most valuable signals is feedback from our community when people mark something as false news.
Today, we only have fact-checking partners in a handful of markets, but are investing in ways to scale these efforts globally.” After the death of Bollywood actress Sridevi, thousands of videos surfaced overnight on YouTube – with most of them trying to sensationalise her death and spreading misinformation about the superstar. While some videos have been taken down, others remain on the platform.
Satya Raghavan, head of entertainment content, YouTube in an interview to ET said that the company has taken down many videos around the actress and the circumstances of her death. “The number of videos that are taken down are five times more now globally.”
YouTube has 225 million monthly active users on its platform and has hundreds and hundreds of hours of content uploaded every minute. “So we empower the viewer to flag off content that they feel may be inappropriate… We have also invested a lot over the couple of years, in machine learning, which can help us recognise a whole lot of this kind of content,” he said.
India is working on a data protection framework which is expected to straighten some ambiguities around data privacy of citizens along with putting liabilities on companies for any leakage or misuse. The law is expected to ready in the next couple of months.