TikTok is investing in a multi-layered trust and safety system combining artificial intelligence, human moderation, local expertise and specialist teams as it manages the growing volume and complexity of content on its platform.
Keagile Makgoba, head of communications at TikTok in sub-Saharan Africa. Image supplied
Speaking on TikTok’s approach to content moderation and user safety, Keagile Makgoba, head of communications at TikTok in sub-Saharan Africa, said the scale of content uploaded to the platform makes technology essential, but that automated systems cannot operate effectively without human oversight and local context.
“We have high volumes of content every single minute,” Makgoba said. “We have to have a multi-layered approach to moderating content.”
For TikTok, the challenge is not only identifying content that breaches its rules but also determining context, language and intent across different markets and communities.
AI and human moderation work together
TikTok uses automated technology to assess content from the point at which it is uploaded, with human moderators and specialist teams providing additional layers of review.
Makgoba said the platform's automated systems begin assessing content during the upload process, before it is distributed more widely.
“The tech already plays a very big role right at the beginning of the stage,” she said.
But technology is complemented by multilingual human moderators who can assess content where automated systems may not fully understand local context.
“If your content has been queued and has to go and get moderated and have someone else with local context, it goes through that process,” Makgoba said.
This is particularly important for a platform operating across different languages, cultures and social contexts.
TikTok also has specialist teams dealing with specific categories of risk, including teams that work directly with law enforcement when situations require escalation.
The model effectively creates several layers of intervention, from automated detection and human review to specialist escalation and external support.
Local context shapes global guidelines
Makgoba said TikTok's Community Guidelines are not developed as a one-size-fits-all global rulebook.
The company consults internal and external experts, including global and regional safety advisory councils, when developing and updating its policies.
“We don't have five people randomly selected, sitting in a room and saying, ‘These are the rules for TikTok,’ and therefore they're going to work across over 100 countries,” she said.
The guidelines are designed around principles including fairness, transparency, inclusion, dignity, privacy and freedom of expression, while also recognising that users' rights need to be balanced against the potential impact of harmful content.
TikTok regularly refreshes its guidelines as technology and online behaviour evolve.
“When I first joined, it was every two years. But with the ever-evolving tech space, it actually happens much quicker,” Makgoba said.
For a global platform, the ability to adapt policies to changing risks and local market conditions has become an important part of maintaining user trust.
Image supplied
Transparency becomes part of platform governance
TikTok is also using greater transparency to address user concerns around content removals, account restrictions and moderation decisions.
The company's Transparency Center publishes information on enforcement activity, including content removed by policy category, while users can access the Community Guidelines through the platform and online.
Makgoba said users should also make greater use of the appeals process when they believe enforcement decisions have been made incorrectly.
She pointed to the frequent use of the term “shadow banned” by creators whose content is no longer being recommended on the 'For You' feed.
“Maybe you weren't shadow banned. You could have made a mistake and you're not recommended on the For You feed,” she said.
TikTok's strike system is intended to give creators greater clarity about potential violations by warning them when they have breached its rules and explaining why.
The company also provides an appeal mechanism, creating another layer between automated enforcement and permanent consequences for an account.
Safety extends beyond content removal
TikTok's approach increasingly incorporates preventative tools that allow users to shape their own experience on the platform.
Users can manage topics, apply keyword filters, control comments and refresh their feeds when their interests change. Creators can also filter or remove comments and manage specific words they do not want appearing in their comment sections.
The company also uses search interventions for areas considered particularly sensitive, including self-harm and elections.
Rather than simply returning search results, TikTok can intervene and direct users towards information or support resources where appropriate.
In South Africa, Makgoba said TikTok works with local organisations and NGOs to inform some of these interventions and connect users with relevant support.
This approach reflects a broader shift in platform safety from simply removing problematic content towards building systems that can intervene earlier and direct users towards appropriate resources.
Different experiences for younger users
TikTok is also applying additional safeguards to younger users.
Users aged between 13 and 17 have a different experience from adult users, with additional controls intended to reflect their age.
Makgoba encouraged parents to use dedicated teen accounts rather than allowing children to use an adult's account.
Parents can also use TikTok's parental controls to manage aspects of a teenager's experience from their own devices.
The company is additionally engaging young users through its Global Youth Council, which contributes feedback on the development of platform features and safety measures. Makgoba said several African teenagers are represented on the council.
Trust and safety as infrastructure
For TikTok, the scale of its platform means trust and safety increasingly operates as a technology and governance function rather than simply a moderation exercise.
The combination of automated detection, human review, specialist escalation, local partnerships, user controls and transparency mechanisms is designed to manage risks while allowing the platform to continue operating at global scale.
At the same time, TikTok is positioning users themselves as part of that safety infrastructure, encouraging them to report violations rather than redistribute harmful content and to use the controls available to manage their experience.
“We need you to play a role in keeping the platform safe,” Makgoba said.
As TikTok expands its creator and commercial ecosystem in South Africa, the ability to maintain user trust while managing content at scale will remain a critical part of the platform's growth strategy.
