Content Monitoring

As more people continue to share information on social platforms, those with bad intentions will continue to find ways to spread malicious and inappropriate content. Terrorists, hate groups, human traffickers, extremists, bullies, and hostile foreign countries have all demonstrated the ability to promote harm and confusion on these platforms. The problem extends beyond public social media platforms. Online gaming is another venue that is significantly affected, primarily through VoIP. While the nature of such adversarial content and the profile of its authors can differ, similar behaviors take place over emails, chats, and on the phone. Current counter-measures are expensive and often have limited effectiveness, notably due to a substantial reliance on historical data to deal with a very dynamic issue that has platform-level specificities. SumUp's Nucleus platform enables a novel approach to tailor and augment detection to improve monitoring scalability and adaptability so you can keep up with the malicious actors and keep your platform safe.

Content Monitoring is a daunting task

Posts at high velocity and volume make identifying objectionable content difficult and expensive.

As nefarious posters get detected and their posts are taken down, they adapt to circumvent the current moderation systems which increases the burden upon the content monitors and the amount of malicious content seen by users.

Content monitors are under high stress due to the nature of the content they keep facing.

Policies widely vary by country and by platform and frequently undergo revisions.

A novel approach

Let users define adversarial content categories based on their platform's data and policies.

Complement automated detection with machine learning algorithms that augment content monitors to learn on-the-fly at large scale.

Adapt and investigate with a multi-faceted transparent approach to keep media platforms safe.

Highly flexible deployment and complete data isolation to satisfy complex regulatory and operating environments.

Sample content monitoring capabilities of Nucleus

Define and update content categories that match your policies

Dynamically search, update, and expand suspicious terminologies

Identify and analyze interactions even without obvious connections

Track relevant conversations and how they evolve over time

Extract novel information that has not been seen before

CASE STUDY

Detection Of Suspicious Tweets For Undisclosed US Agencies

Challenge

Challenge

  • Threat analysis of 20 million tweets across 8 languages, covering 5 hours of content during the US-NK summit
Solution

Solution

  • Leverage SumUp APIs to identify potentially threatening tweets and their authors
  • Operate at the conversation level and single-post level
  • Uncover potential threats without a-priori and with targeted analysis
SumUp Advantage

SumUp Advantage

  • Accelerate threat assessment from two weeks to three hours
  • Retain a high-degree of control, flexibility, and traceability