On-demand webinar: navigating the challenges to AI success
Resources
Case Studies

Creating a Safer Web Experience

Trust Lab: Making Social Media Safer Through Sentiment Analysis
November 2, 2022
Share

The Project: Creating a Safer Web Experience for All

Trust Lab, a trusted company founded by senior Trust & Safety executives from Google, YouTube, Reddit and TikTok, is ensuring the internet remains a safe place for all. They are doing this through sentiment analysis to see what real internet users think about various content that is shared on the world wide web. Since March 2020 people are spending 45% more time on social media globally according to the Enough Organization.

The Challenge: People’s Perspectives are Unique

Social media is consumed by people worldwide and everyone who uses it comes from various backgrounds and has different viewpoints on content. Creating rules for a platform to make everyone feel safe while using it is difficult. What one person finds offensive; another finds ok. Furthermore, policies can vary significantly between platforms. Which means some platforms condone more controversial topics like nudity while others do not. The data needed to support what policies are best served for a given platform come from getting the opinions of users across the globe. Trust Lab determined the best way to get everyone’s opinions is through user sentiment data based off their opinions on various user generated content (UGC).

The Solution: Clearly Identified Trust and Safety Plan

Leveraging our Appen Data Annotation Platform (ADAP), datasets containing UGC content covering all topics imaginable were added. Our expansive data labelling options within the platform allowed the organization to determine the exact types of sentiment they wanted feedback on for each piece of UGC that was to be reviewed. The company was able to receive detailed sentiment analysis data that let them see how much of the content was bad, as determined by the team of worldwide reviewers.

Trust Lab uses the data to provide metrics to the various platforms on the internet to help them understand what type of content is on their platform, how it compares to other platforms, and how the public feels about the content. Those platforms are then able to adjust their trust and safety plans or remove the offensive content.

The Result: A Safer Tomorrow

Using Appen’s platform, contributors made millions of judgements. The sentiment data received is one input used to create a detailed trust and safety enterprise system with recommendations for their partners on what needed to be changed and why, identifying high risk and unsafe content, accounts and transactions at scale. The information from Appen helps to add additional context to Trust Lab’s trust graph products.

Companies with detailed trust and safety plans convey to their users that they genuinely care for their wellbeing and want them to be long time users of the app. Users who feel safe, stay on those platforms. With the help of data annotation these platforms now have the capability to enforce their plans creating a safer tomorrow.

“Appen’s expertise and ability to scale has provided Trust Lab with the ability to provide user sentiment metrics to our clients to ensure their platforms are safe for everyone.” shared Nick Miller, Senior Data Scientist at Trust Lab.