Seoul, South Korea – October 29, 2026 – Metacrowd Corp., a deep-tech artificial intelligence detection and authenticity verification enterprise, today announced that it will present its comprehensive multimodal security suite—comprising MetaShield, VoiceShield, and TextShield—at Web Summit 2026 in Lisbon, Portugal, targeting strategic collaborations with European financial institutions, telecommunications carriers, and legal organizations.
A New Paradigm in AI Verification and Trust Infrastructure
Founded in February 2022 and headquartered in Seoul, Metacrowd Corp. operates under the core mission of “Building Trust in AI” by engineering technologies that detect synthetic and manipulated media across audio, video, and textual assets. The rapid rise of generative AI has lowered barriers to creating sophisticated voice clones, deepfake video feeds, and synthetic documentation, creating substantial operational vulnerabilities for corporate institutions and public infrastructure.
To address these systemic threats, Metacrowd developed a modular detection ecosystem that identifies synthetic signals across separate communication vectors. Recognized for its technical approach, Metacrowd was selected for South Korea’s prestigious TIPS program, supported by the Ministry of SMEs and Startups, as well as the Google for Startups Accelerator.
Core Technologies and Certified Product Suite
At Web Summit 2026, Metacrowd will showcase its three flagship platforms, each engineered to address specific synthetic media threats:
• MetaShield: A real-time deepfake video and image detection solution capable of analyzing live video streams within two to four seconds. The platform conducts itemized forensic evaluations of facial boundary-blending artifacts, abnormal blink rates, and lip-voice synchronization mismatches, generating frame-by-frame forensic reports designed to serve as formal evidence in legal proceedings and investigations.
• VoiceShield: A real-time call security application that verifies whether a caller’s voice is AI-synthesized while simultaneously evaluating conversation patterns for social engineering indicators. Analyzing subtle acoustic markers such as breathing cadence, intonation irregularities, and anomalous frequency distributions, VoiceShield provides verifiable explanations for its risk assessments. Engineered with a plug-in architecture, the underlying model accommodates new voice-synthesis patterns and additional languages—such as English and Arabic—without requiring core system re-engineering.
• TextShield: An on-premises document analysis platform that inspects academic manuscripts, legal submissions, and corporate reports paragraph by paragraph. The platform evaluates whether generative AI was utilized and determines the exact degree of intervention, ranging from minor grammatical polishing to full synthetic generation, all while retaining documents entirely within local enterprise servers to prevent data exposure.
Certified Benchmarks and Commercial Traction
Metacrowd’s platforms have undergone extensive third-party testing by accredited institutions, establishing commercial references across the legal, academic, and public sectors.
In formal evaluations conducted by the Korea Testing & Certification Institute (KTC) using international public benchmark datasets, MetaShield recorded a 98.7% deepfake detection accuracy rate, yielding a false-positive rate of 0.37% on authentic media and a false-negative rate of 2.75%. Based on these certified results, leading South Korean law firms utilize MetaShield to analyze forensic evidence in deepfake litigation, while legal practices in the United States are currently evaluating institutional adoption.
VoiceShield completed KTC operational testing under real call conditions in 2026, recording an error rate of 2.78% on fraudulent calls and a false-positive rate of 1.39% on standard calls, outperforming four competing commercial applications tested across identical parameters. In comparative evaluations against US-based Reality Defender, VoiceShield demonstrated higher detection accuracy across all six comparative test groups. Following its initial rollout in South Korea, the English version launched in Singapore in March 2026 to validate detection across mixed linguistic environments containing English, Mandarin, Malay, and Tamil. In a municipal pilot conducted with Gwangju Metropolitan City targeting elderly demographics, the proportion of users terminating suspicious fraudulent calls increased from 28% to 52%.
In academic governance, TextShield was officially adopted as the primary manuscript screening platform by the peer-reviewed journal AI Humanities Research, validating its bilingual processing capabilities across Korean and English texts.
“Generative tools have evolved to the point where human perception can no longer distinguish authentic interactions from synthetic fraud,” said Hyungjin Lukas Kim, CEO of Metacrowd Corp. “Protecting institutions requires multimodal AI that operates across voice, video, and text simultaneously, delivering forensic-grade verdicts in seconds. Web Summit 2026 provides the optimal venue to demonstrate our certified detection accuracy to European banking groups, telecom carriers, and regulatory authorities navigating compliance with emerging AI governance frameworks.”
Global Expansion and Strategic Roadmap
Metacrowd is utilizing its presence at Web Summit 2026 to initiate its commercial entry into the European Union, where enforcement of the EU AI Act and mandatory synthetic content labeling mandates are accelerating demand for certified detection tools. Throughout the event, Metacrowd will operate interactive live demonstrations at its booth, featuring real-time deepfake analysis of live video calls, simulated voice-phishing detection scenarios in English, and instant document inspection.
The company is actively seeking partnerships with European financial institutions seeking to embed call verification into mobile banking applications, telecommunications operators evaluating native device integration, and litigation firms requiring evidentiary video forensics.
Looking ahead, Metacrowd plans to commercialize VoiceShield in North America through local corporate partnerships, initiate Middle Eastern beta testing utilizing its Arabic detection model in Saudi Arabia and the UAE, and unify its specialized technologies into an integrated enterprise AI security suite by 2028.
About Metacrowd Corp.
Founded in February 2022, Metacrowd Corp. is an artificial intelligence detection and authenticity verification company dedicated to building trust in AI. The company develops multimodal technologies that identify synthetic media across voice (VoiceShield), video and images (MetaShield), and text (TextShield). Headquartered in Seoul, Metacrowd serves clients across the legal, financial, and academic sectors with high-precision, certified verification tools designed for real-time and forensic applications.
Press Contact
Hyungjin Lukas Kim
CEO
Metacrowd Corp.
Email: lukas@metacrowd.ai
Phone: +82-2-300-0743 / +82-10-9161-7178
Website: www.metacrowd.ai
Filed in . Read more about AI (Artificial Intelligence), Cybersecurity, Korea, Korean Startups, Security, Web Summit and Web Summit 2026.