Safety @ Koo

Koo’s core mission is to provide users a safe and trusted micro blogging platform in a language of their choice. In order to provide our users with a wholesome community and meaningful engagement we take a number of steps to keep the platform safe. These steps can be broadly classified into three categories:

1. Data Security & Privacy
2. Handling Misinformation & Fake News
3. Handling Hate Speech

Data Privacy & Security

Privacy:

Koo has utmost respect for user rights to data that belongs to the users. We are transparent about what we collect, the purpose the data will be used for and give users the control of what they want to share and how they want to deal with the platform.

Consistent with the requirements of local laws (The Lei Geral de Proteção de Dados in Brazil is an example) Koo is committed to ensuring that we protect the rights of freedom and privacy which have been granted under the laws.

No personal data or sensitive data is collected without consent and the purpose for collection is clearly mentioned in our Privacy Policy. All personally identifiable information is encrypted, is inaccessible to everyone and is stored in the host continent itself.

The policy itself is available in English and native languages and is easily accessible from within the App and from our website. In line with the requirements of free access under the LGPD, the Koo Privacy Policy has information about (i) the categories of data that we collect and the purpose of that collection; (ii) the contact information of the Data Protection Officer (Article 41); (iii) the contact information of the Grievance Officer; (iv) why and under conditions data may be shared with agents or service providers, and (v) the most fundamental rights of users listed in Article 18 viz

1. Access
2. Rectification
3. Elimination
4. Objection
5. Portability
6. Consequences of not providing consent, and
7. Process and forms for submitting requests for cancellation

Security

Koo compiles with ISO 2700: 2013 standards while collecting, storing, sorting and handling data. User data is encrypted at rest and in transit. Data of Brazilian users is encrypted and stored within Brazil (Sao Paulo).

Our employees undergo periodic training on data security under the supervision of a Data Protection Officer and we implement strict access controls to ensure that data of users is kept safe. We also implement round the clock monitoring of our systems to detect and repel any attacks using Cloudflare and other industry standard products.

In addition, all Koo service providers are given information relevant only to the task they perform, on a need to know basis, and in an anonymised and encrypted form. These Service Providers must also follow our B2B Privacy Policy and are subject to audit by Koo Security team from time to time.

Handling Misinformation & Fake News

Koo Community Guidelines specifically prohibit sharing of false, misleading or unverified information. While this sounds simple, handling misinformation is one of the most difficult problems today. It is important to understand that misinformation does not start with the social media company. Misinformation starts with certain users posting something that they either know to be false or misleading, or unknowingly posting material that is false, but they believe to be true.

Koo’s approach to handling misinformation is based on several layers.

1. Koo provides its users the means to voluntarily self verify themselves as real humans who have a genuine voice on the internet and want to be seen. We believe that users are more likely to trust accounts which declare themselves to be “not bots”. We also believe that such self declared accounts and Eminent users are more likely to act with a high sense of responsibility while dealing with information. To this end, self-verification and Eminence are free on Koo and will always be free.
2. Any user on the platform can report content that they believe to be false or inaccurate using the in-app reporting mechanisms. This content is flagged to our trust and safety team for review. Users have also been empowered with tools to directly reach out to fact checkers. We constantly keep curating and listing fact checkers on our website and users can interact with them without any interference from the platform.
3. The third component of our approach is to proactively detect, flag and if found egregious, remove false content. Our algorithms look for spikes related to events, trending hashtags and abnormal actions by user handles. Based on the severity and frequency of these spikes, we remove content which is outright false while in other cases we may take a combination of steps, including, slowing circulation, elevating/surfacing content from expert/Eminent accounts or overlaying with warnings.

Handling Hate Speech

Koo subscribes to the definition of hate speech as stated in the UN Strategy and Plan of Action on Hate Speech, to mean any kind of communication whether in speech, writing or behaviour, that attacks or uses pejorative/contemptous or discriminatory language referencing a person or a group on the basis of their identity for example religion, ethnicity, nationality, race, colour, descent, gender or other identity factors. We believe the UN’s definition gives enough direction to be applied to all situations.

While evaluating whether certain content is hateful or not, we look at a mix of:

(i) the stature of the person making the statement — whether the speaker is a person of prominence or has a large follower base

(ii) proximity (physical or social) of the maker to the subject of the statement — is this just another far away commentator or a person who is close, influential and can create a real impact on the subjects

(iii) the context in which the statement is made — is there a burning issue which gets inflamed further, and

(iv) any indications of actual physical threat to the persons who are subjected to the statements. We also try to understand if the statement is legal in the country where the maker of the statement is situated.

The management of hate speech is triangulated on meaning, severity and frequency. Prioritization is given to the most consequential cases first, e.g. if there is a call for physical harm.

Another principle we follow is to manage the content first (and not the creator). If a user is flagged to be a repeat and persistent violator, the user is assisted with reminders. These reminders are sent at thresholds linked to frequency and severity of violations.

In extreme cases we implement our version of digital law enforcement (also similar to offline law enforcement). We may take steps to restrict visibility to others or of others; ability to notify others or receive notifications from others or restrict ability to comment, like or Re-Koo or simply being blocked from the platform.

At Koo we believe that truthful and respectful communication and communities are based on everyday common sense rules that are already practiced in the offline world. We will continue to ensure that speech remains secure from attack, free, honest, respectful and meaningful for everyone.