Google and Jigsaw stay true to its commitment to fight internet trolls by open sourcing a new API called Perspective. Developers will now be able to access a machine learning anti-harassment tool that could ward off the proliferation of abusive speech online. Critics say that this new tool may be a form of censorship as publishers may opt to delete perceived toxic comments. The new API is actually a way to combat censorship by allowing developers to enable comments knowing that a tool is present to detect toxic language and eventually pave the way for a return to healthy debates.

Jigsaw, an offshoot of Google, along with the Google Counter Abuse Technology Team has just open-sourced Perspective, an API that provides developers with anti-harassment tools. The new code utilizes machine learning to identify offensive remarks like insults, harassment or any kind of abusive language. Google Perspective is actually part of the much larger Conversation Artificial Intelligence (AI) Initiative taken by Jigsaw as it declares war on internet trolls since September last year.

Google Perspective can easily measure the level of "toxicity" in every sentence written. It is more accurate than blacklisting keywords and acts quicker than a human moderator, Wired reported. The Conversation AI can facilitate a return to civility with a troll-fee discussion online. It will also filter abusive comments or as its critics would say "sanitize public discussions." It is an internet antitoxin that will no longer allow toxic comments online, but it has not been well received by some sectors in the online community.

Free speech groups have voiced out their fears that the Google Perspective may be one way of imposing censorship. It is a machine after all that could still perceived comments as toxic when they are not. Moreover, publishers without the extra resources to hire a moderator may opt to delete the comments without further investigating when the Conversation AI rates some comments as toxic. However, Jared Cohen, president and founder of Jigsaw, thinks otherwise and believes that the Perspective API will actually reduce censorship.

Cohen adds that the default position now being taken by developers and publishers are to do away with comments altogether. This is worse than censorship for it does not promote healthy conversations. Google Perspective can actually change that by removing toxicity and engaging people to talk and listen online even though they may not always agree.

For Cohen, tools like Google Perspective is the kind of technology that may bring back the days of respectful and informed comments from readers. The same kind of comments seen in the New York Times which employs more than half a dozen team of moderators, according to The Atlantic. With the new API, websites now have a more affordable way of warding off internet trolls. In addition, Jigsaw hopes to improve the Perspective API beyond its focus on toxicity like identifying when a comment is off-topic as starter.

Read more
Tags Google, Machine Learning