i am safe online

A cloud service that blocks offensive content and online harassment.

## online harassment

We all know that online abuse is a real and common problem.

Research and surveys show that it affects all kinds of people and can have serious negative effects if it’s left unattended.

## social media platforms can do better

Recently the problem was addressed by the major social platforms. Everyone seems to be working on the solution, but it’s not as effective as it should be.

Instagram announced the big change by introducing an automated tool, which hides notifications that contain certain words… However, it turned out that it is only available to celebrities. So unless you’re Taylor Swift, you’re still one on one with an abuser.

It just doesn't seem to be a priority. Twitter showed that they could block content they deem inappropriate really fast. But they only use it when someone’s financial interest is involved.

Facebook’s system is only activated by a user reporting abuse, and it’s altogether slow, complicated and inefficient.

## i am safe online

i am safe online offers an easy to use service that can be integrated into any platform.

As a fundament to our service, we provide an API that gives developers access to a raw classifier just one request away. This way they can use the output according to their needs.

Besides, we will develop plugins to block abusive content for popular blogging platforms that can be used by individuals - and it is going to be FREE.

For the enterprise and businesses, we will provide a flexible subscription model.

We think that many governmental institutions will be interested in our service such as libraries and schools, too.

## under the hood

To classify the content we’re a building a machine learning model that uses a large dataset of anonymised abusive and nonabusive comments from Reddit & other social platforms.

We’re using natural language processing and a recurrent neural network that allows us to look at the comment as a whole rather than treating it as a set of separate words.

In this way, the algorithm learns the traits specific to abusive content and picks the meaning of it.

Our service is built on a scalable platform from the very begging and uses latest technology stack, which makes it secure, super fast and ready for a high traffic load.

**everyone deserves to be safe & happy**

Transcript

1. @kakauandme & @thekoshkina i am safe online
2. @kakauandme & @thekoshkina 52% of young people report being cyberbullied 95% of teens who witnessed bullying on social media report that others, like them, have ignored the behaviour only 5% of internet users who witnessed online abuse have reported it to the law enforcement 2 online harassment
3. Cyber-bullying has long term effects on mental health of a victim. —Professor Sheryl Hemphill, Melbourne University 3
4. @kakauandme & @thekoshkina social media platforms can do better 4
5. @kakauandme & @thekoshkina 5
6. @kakauandme & @thekoshkina 6
7. @kakauandme & @thekoshkina 7
8. @kakauandme & @thekoshkina 8
9. @kakauandme & @thekoshkina 9
10. @kakauandme & @thekoshkina i am safe online 10
11. @kakauandme & @thekoshkina simple API for everyone to use out of the box harassment prevention tools for popular platforms free for the personal use flexible subscription model for enterprises, business and governmental organisations i am safe online 11
12. @kakauandme & @thekoshkina verified dataset off 33,641+ abusive comments deep recurrent neural network for harassment classification scalable and powerful engine hosted on Google Cloud Platform under the hood 12
13. @kakauandme & @thekoshkina 13 everyone deserves to be safe & happy