Artificial Intelligence is learning how to take down internet bullies by being trained to recognise patterns of abuse to curb trolling.
On the 6th of August the World Wide Web will celebrate its 26th birthday. The introduction of the Internet has dramatically influenced the last quarter-century, for better and for worse.
While this great invention has provided us with many benefits, we have also seen the rise of many illegal and dark activities on the Internet, from abusive trolls to recruitment drives for terrorism.
One of the giants within the tech industry, Alphabet, is fighting to tackle the mammoth task of purifying the Internet’s darkest corners. The company, formerly a part of Google, has developed Jigsaw which focuses on utilizing new methods to confront the troubles of the Internet.
Read More: How Tech Companies are Combating Terrorist Propaganda
The company is combating issues such as online channels used by terrorist organizations, e.g. ISIS, to brainwash potential recruits. To do this the project has sent researchers around the world to interview defective terrorists in order to understand their psyche. With a better understanding of what appeals to potential recruits and how they are recruited, Jigsaw can counteract these messages and channels with positive information, disrupting previously unchallenged streams of brainwashing content.
Abusive comments and Internet trolls are also in the cross hairs for this ambitious project. Jigsaw is training Artificial Intelligence (AI) to recognise patterns of abuse to take down online bullies. The aim of this task is to build a moderation system, using AI and machine learning, allowing for better comment moderation on websites.
Read More: Google’s new Perspective API can help you not sound like a jerk while commenting
During an interview with Mashable, Yasmin Green, director of research and development at Jigsaw, explained that machine learning is an interesting process because you don’t give AI rules, you give it patterns and it learns.
This method is not flawless, but they are working to ensure the system is not biased. For example, the algorithm once consistently flagged the word “gay” as hateful, due to the fact that many comments used to train the system had negative connotations. To counteract this, they added news stories that included the word “gay” into the system.
Read More: How Microsoft’s AI chatbot ‘Tay’ turned incestuous Nazi accurately reflects our society
Jigsaw is fortunate enough to have the funding of Alphabet enabling it to carry out this benevolent research. However, nasty comments and intentionally damaging posts are also a big problem for corporations with a high social media presence.
One company that is looking to solve this issue for businesses is BrandBastion. Much like Jigsaw, this Finnish company is working to provide a service using AI to minimise negative and potentially harmful comments for companies and their social media sites.
Read More: Instagram Partners With AI Platform to Manage, Improve Social Engagements
“There is a lot of great stuff with social media,” explains the CEO, Jenny Wolfram, to Good News Finland. “It enables people to connect, also with the brands that they love. But there is also a dark side.”
“There is a lot of engagement that is not in the best interest of people and fans that are trying to communicate with their brand,” Wolfram explains. “This could be spam, phishing attempts, threats, malware, pornography or free gaming gems being spread.”
Despite its many great offerings, it is true the internet can be a very dark place at times. Internet trolls and their obsession with causing havoc are not likely to disappear anytime soon.
However, using the right technology, we can help curb their impact and influence across the web. As AI becomes smarter and more informed, we can hopefully create a safe space on the internet for people to socialize and exchange ideas, without keyboard warriors in their mum’s basements spitting out insults for personal kicks.