Social platforms face risks that can move across accounts, conversations and platforms. Fraud, fake identities, online rumors and illicit online operations may appear through private messages, external links or patterns of activity that are difficult to spot from a single post. The earlier those signals are identified, the better chance a platform has to warn users before they suffer actual harm.
Soul App, a Chinese AI social platform, has spent recent years strengthening its safety systems through AI-based risk detection, user alerts, anti-fraud education and cooperation with public security authorities. Its work covers content, accounts, devices and interactions, with public security authorities involved when cases require further investigation or law-enforcement action.
Detecting Risks Earlier
Soul App continues to refine its AI models and introduce targeted strategies, extending detection from individual pieces of content to accounts, behavior and interaction chains. A private conversation, an image or an external link may look on its own. The problem can become clearer only when these signals are considered alongside account activity and patterns of user behavior.
This kind of detection gives Soul App more opportunities to users and intervene before suspicious behavior leads to substantial harm.
In the first quarter of 2026, Soul App’s automated review systems directly blocked 4.77 million pieces of content that violated platform rules and took action against 162,000 accounts. In the second quarter, the systems blocked another 4 million pieces of content that violated platform rules, handled 7.26 million related comments and took action against 140,000 accounts.
After Soul App introduced a system for identifying large-scale illicit online activity and cheating, the handling rate related to risky content rose from around 30% to 60%. Its anti-fraud Agent has reached a detection accuracy of more than 93%.
Working With Public Security Authorities
In recent years, Soul App has worked with public security authorities on fraud prevention, illicit online operations online rumors and cybersecurity. Within legal and compliance requirements, Soul App identifies suspected risks through its internal safety systems and assesses potential leads. For higher-risk cases involving fraud, illicit operations and rumors, it provides technical and platform support. Public security authorities can investigate suspected criminal activity and take action under the law.
Soul App’s received the "Best Supporting Unit of the Year" award from the cybersecurity unit of the Pudong Branch of the Shanghai Municipal Public Security Bureau for its support of efforts against online rumors in 2024. In the first quarter of 2026, Soul App took part in several initiatives targeting fraud and illicit online operations and submitted more than 50 valid leads.
Soul App also brings fraud-prevention education to users through safety outreach exchanges with public security authorities and explanations of platform rules. Soul App has invited police officers specializing in fraud prevention to join audio chat rooms on the platform, where they explain common scam tactics and answer users’ questions.
Soul App’s safety work starts with protecting its own users and the environment in which they interact. AI helps identify earlier, while alerts, platform action and external cooperation provide different ways to respond when those risks emerge.
As more social interaction takes place online, user safety depends on more than removing harmful content after the fact. Platforms need ways to detect risks earlier, warn users and cases that exceed their own authority. Soul App is trying to turn the safety capabilities built within its platform into part of a broader collaborative effort involving public security authorities, regulators, industry organizations and users, the aim of creating a safer and more orderly online environment.
Copyright © 2026 California Business Journal. All Rights Reserved.
For California Business Journal Disclaimers, go to https://calbizjournal.com/terms-conditions/.
