Because online interactions now both define and shape social reality, the question of safety and authenticity has become central to every digital platform’s existence. This spurred the CEO of Soul App, Zhang Lu, to make AI a part of the digital safety equation on the platform.

The fact is that social networking apps like Soul are more than mere spaces for self-expression. They are in effect new age cultural mirrors that reflect the complexities, vulnerabilities, and ethical responsibilities of modern connectivity. As such, these platforms have a responsibility to create safe, secure, and generally positive ecosystems for their users.

Soul Zhang Lu quickly realized that cutting-edge technologies like artificial intelligence could play a crucial role in adhering to the platform’s ethos of offering a secure space for emotional expression and meaningful connections. Soul, which enjoys a stellar fan following among China’s Gen Z, also wanted to maintain transparency when it came to the safety and security protocols used on the platform.

So, the social networking app regularly puts out security reports detailing how the company’s fusion of technology, governance, and education is helping to create a safe and welcoming digital arena for its members. The Q3 2025 Ecosystem Security report from Soul Zhang Lu’s team outlines the platform’s most recent efforts in this direction and discusses the results they have yielded.

AI-Driven Risk Control System

In order to build a smarter defense strategy, the social networking app has turned AI into the guardian of social integrity by placing AI-powered risk control at the heart of the platform’s digital safety framework. Over the past few years, social networking platforms worldwide have relied on machine learning models to detect harmful behaviors such as fraud, cyberbullying, and misinformation.

Soul’s cybersecurity experts have fine-tuned this approach by integrating AI large models across multiple layers of platform activity. This helps to identify and intercept risks before they turn into an issue for users.

Soul Zhang Lu’s team has tasked proprietary AI models with enhancing precision and response efficiency. Many of these models were upgraded in 2025, and these enhancements have led to a significant improvement in performance. For instance:

• The optimization of Soul’s AI image anti-fraud model resulted in an 80% reduction in false positives, while increasing coverage accuracy to over 90%.
• After the upgrade, the platform’s persona large model, which is designed to detect high-risk accounts posing as deceptive identities, saw its detection accuracy rise from under 50% to above 70%.

Since the emphasis is on identifying potential fraudsters early, Soul Zhang Lu’s team is able to prevent malicious users from infiltrating its social graph. In addition to giving itself the technological edge needed to keep antisocial elements in check, the social networking platform continues to collaborate closely with law enforcement and regulatory agencies. In Q3, Soul submitted over 100 leads related to harmful content and participated in multiple cooperative investigations, contributing to national efforts against illegal online activity.

Human-Machine Collaborative Content Governance Model

Despite the impressive performance of the platform’s AI models, Soul Zhang Lu gauged that a holistic model, which integrates automation with human insight, is what’s needed to maintain both safety and community vibrancy. So, instead of creating an AI-only governance model, Soul’s team opted for an AI-assisted governance model.

This strategy has proved astoundingly effective at addressing the core and perennial challenge faced by social platforms of protecting users from harm without suppressing creativity or open dialogue. Soul’s report revealed that in the third quarter, the platform:

• Intercepted 4.39 million pieces of violating content.
• Processed over 15 million comments and replies for moderation.
• Sanctioned 306,000 violating accounts.

Furthermore, the platform also boasts an automatic system that blacklists devices linked to multiple infractions. In a nutshell, the report clarifies that what Soul Zhang Lu has is a scalable trust infrastructure capable of sustaining user safety and confidence, without compromising on user engagement.

Soul’s group chats and community spaces also incorporate extensive AI-driven monitoring. As per the report, on average, the system blocked 5,600 non-compliant voice messages and 26,000 violating text messages per day. These measures have helped to ensure that real-time social spaces remain civil and inclusive.

Empowering Users: Building a Self-Defense Network

Moreover, Soul’s advanced text and audio pattern recognition models are essentially intelligent shields, designed to learn from evolving user behavior. But instead of handing over complete control to advanced technology, Soul Zhang Lu believes in empowering users through the tools and awareness that they need to protect themselves in online environments.

To this end, Soul launched over 20 new private chat security features aimed at enhancing user control and autonomy. One of these is the platform’s “Self-Visible Message” policy. Thus far, this security feature has prevented over 43,000 users from receiving unwanted or harassing messages. Also, accounts with repeated infractions faced an average of 80,000 private chat restrictions per day.

In the same vein, another freshly introduced feature from Soul Zhang Lu’s team, known as “Safety Reminder Pop-Ups,” is designed to proactively notify users if they may be targets of harassment, encouraging them to act before escalation occurs.

Special Crackdown: Focusing on Cyberbullying and Youth Protection

The social networking platform has taken a similar, comprehensive stance on the issue of cyberbullying. The report states that over the last quarter, the platform protected 3,127 users daily from cyberbullying and harassment, handled 2,100 daily reports of uncivil behavior, and issued 215,000 reminders per day encouraging positive communication.

Simply put, this dual mechanism that amalgamates risk identification and user education is at the core of Soul’s governance philosophy. Furthermore, Soul Zhang Lu remains committed to youth safety. In keeping with this stand, the platform enhanced its age and identity recognition models by 200%.

Conclusion: Toward a Secure Social Ecosystem of Shared Governance and Shared Benefits

In light of these forward-thinking security and safety measures, it can be concluded that the app is making a concerted effort to create a “clean, healthy and positive” digital space for the platform’s young users. Also, by turning platform safety into a collective responsibility shared by all users, Soul Zhang Lu’s team is expanding the reach and scope of the app’s security measures while amplifying user engagement and strengthening the collective consciousness of digital citizenship.