In the World in You x Org Interview Series, we shine a spotlight on the trailblazing impact organizations and visionary entrepreneurs dedicated to forging a brighter, more sustainable society.

In July 2025, we welcomed Tomáš Halász, Co-Founder and CEO of TrollWall AI, a Slovakia-based startup at the forefront of combating online hate.

Enjoy the full conversation on Youtube.

Interviewer: Mio Yamamoto, Article: Kenichi Shibuya and Mio Yamamoto

1. Background of TrollWall AI's Establishment and Founder Tomáš's Experience

As a social media manager, Tomáš frequently encountered a large volume of hateful comments which became a significant mental burden. The urgent need for a solution intensified, especially after comments surged following the invasion of Ukraine. With no suitable tools on the market, he began developing an AI solution with an IT expert friend. What started as a project born from personal need evolved into a full-fledged business in response to demand from other NGOs, media outlets, and corporations. TrollWall AI began from a pressing issue on the ground and Tomáš's personal conviction.

2. Evolution from Harmful Comment Management to Fostering Healthy Communities

Initially, TrollWall AI specialized in hiding hateful comments, such as hate speech and incitement to violence. However, its functionalities have since expanded significantly. It now evaluates comment sentiment (positive, negative, neutral) and suggests recommended actions to social media managers. For instance, for positive comments, it recommends actions like "liking" or replying, even drafting response texts using AI, thereby efficiently boosting positive engagement within communities. This has evolved the tool beyond mere removal of harmful information to one that actively promotes healthy discourse.

3. Extensive Client Base and Free Support Initiatives

TrollWall AI serves a diverse range of clients, currently utilized by clients across Europe and Latin America. Its clientele spans public-facing organizations like politicians, military bodies, and news media, to professional sports clubs and corporations (including telecom companies, banks, and advertising agencies). Notably, sports clubs adopt the service to maintain passionate discussions while eliminating hate speech, and corporations integrate it to protect their brand image. Furthermore, TrollWall AI operates on a unique business model that aims to balance profitability with social contribution, and as part of this commitment, it provides its service free of charge to smaller NGOs, currently supporting over 50 organizations.

4. The Difference Between Moderation and Censorship: Hate Comments Are Not Opinions

The interview strongly emphasized the clear distinction between moderation and censorship. Moderation involves managing harmful conduct to uphold the rules of a specific community, without infringing on freedom of speech. Tomáš unequivocally states that "hate comments are not opinions" and explains that while differences of opinion are not moderated, comments targeting specific groups with undue disparagement or dehumanization, as well as incitement to violence, are always removed. He argues that moderation is essential for preventing noisy minorities from dominating discussions and for creating a democratic space where everyone can participate safely.

5. Unique AI Training Methodology Supporting High Accuracy

TrollWall AI's model is built upon open-source NLP models, trained with proprietary data and methodologies. Central to this is a rigorous annotation process performed by local linguistic experts. Each comment is independently reviewed by multiple experts, and only when all agree on its categorization (e.g., "hate" or "opinion") is it used for training. This method ensures the AI model accurately understands linguistic nuances and local contexts, leading to high-quality results. Continuous human auditing and retraining allow the system to adapt to evolving language trends.

6. The Value of the World Summit Award (WSA)

TrollWall AI was honored with the World Summit Award (WSA) in 2024. Tomáš highlighted that the true value of this award lies not merely in the recognition itself, but in the invaluable networking and exchange opportunities it provided with social entrepreneurs from across the globe. The WSA conference fostered the sharing of knowledge and experiences across borders, significantly aiding TrollWall AI's international expansion. This underscores the crucial role of international platforms for mission-driven organizations.

A Message from TrollWall AI to Our Japanese Readers

Hello to our Japanese readers! TrollWall AI provides an AI moderation service that protects communities from hate speech and harmful comments on social media. We currently support nine languages, primarily across Europe, and are trusted by many corporations, news media, political organizations, and NGOs.
Our goal is to remove hate from online spaces and create a healthier digital society where everyone can engage in constructive discussions with confidence.
We deeply hope to bring this valuable service to the people of Japan in the future. We believe that developing a Japanese language model is achievable, and as a first step, we are actively seeking partnerships with major Japanese companies and organizations. If you are interested in TrollWall AI's services, please do not hesitate to contact us. We eagerly await hearing from you.

Interview Afterword

Through this interview with Tomáš from TrollWall AI, we were once again reminded of the critical importance of maintaining healthy online communities. Two points, in particular, left a lasting impression.
"Hate Speech Is Not an Opinion":
Tomáš emphasized a clear distinction between mere criticism or differing opinions and hate speech that attacks individuals and fosters division. TrollWall AI, while respecting diversity of opinion, holds a strong conviction never to tolerate "hate" such as incitement to violence or dehumanization. We believe this perspective is essential for ensuring healthy freedom of speech in digital spaces.
The Value Gained from the World Summit Award (WSA) Recognition:
TrollWall AI received the World Summit Award (WSA) in 2024. Tomáš shared that the greatest value derived from this award was the opportunity for networking and exchange with social entrepreneurs from around the world. This experience facilitated the sharing of knowledge and experiences across borders, significantly aiding their international expansion, highlighting the importance of international platforms for mission-driven organizations. Finally, as Tomáš himself strongly encouraged, for all Japanese entrepreneurs with a social mission who are working to create a better world through digital technology, we highly recommend considering applying for the World Summit Award (WSA). This experience will undoubtedly be a wonderful opportunity to further expand your activities.

Tomáš Halász, Co-Founder and CEO of TrollWall AI

Tomáš Halász) さん、TrollWall AIの共同創設者、CEO：

TrollWall AI

Tomáš Halász is the Co-Founder and CEO of TrollWall AI, a Slovakia-based startup that developed an AI-powered moderation tool with human-level accuracy. TrollWall AI was born out of the urgent need to address the growing wave of hateful comments on social media. As a community manager and activist, Tomáš saw firsthand the harm caused by toxic online behavior and recognized the need for a tool that could effectively filter harmful content.

Today, TrollWall AI protects hundreds of social media profiles across seven countries from reputational risks by automatically detecting and hiding hate speech and harmful comments. The tool is used by a wide range of users—including newspapers, PR agencies, companies like O2, NGOs, the IIHF Ice Hockey Championship, and even national defense institutions. Prominent public figures such as Czech President Petr Pavel and Slovak President Zuzana Čaputová are also among its users.

◆Contact Us