{"id":2624,"date":"2026-09-23T14:07:56","date_gmt":"2026-09-23T14:07:56","guid":{"rendered":"https:\/\/www.gsma.com\/solutions-and-impact\/technologies\/security\/scams\/?post_type=gsma_theme_study&p=2624"},"modified":"2026-09-25T06:40:25","modified_gmt":"2026-09-25T06:40:25","slug":"indosat-ai-anti-scam-protection","status":"publish","type":"gsma_theme_study","link":"https:\/\/www.gsma.com\/solutions-and-impact\/technologies\/security\/scams\/gsma_study\/indosat-ai-anti-scam-protection\/","title":{"rendered":"Indosat Ooredoo Hutchison: AI anti-scam protection"},"content":{"rendered":"\n

Indosat Ooredoo Hutchison<\/a> (IOH) built an AI anti-scam protection system to counter the growing threat of scams across Indonesia. SMS and voice calls remain the two main channels through which scammers reach mobile users. Global digital scam losses exceeded USD 1 trillion in 2024, according to the Global Anti-Scam Alliance. Indonesia has likewise experienced a significant rise in scam activity, highlighting the need for more scalable and proactive defences.<\/p>\n\n\n\n

Mobile operators must detect threats at the speed and scale that manual review cannot match. They also need to comply with strict national data protection requirements. Consequently, IOH\u2019s system addresses both challenges. It runs entirely within Indonesia, retains no content from any call or message, and requires no human intervention.<\/p>\n\n\n\n

\n\n\n\n

Solution<\/h2>\n\n\n\n

How AI anti-scam protection works across the network<\/h3>\n\n\n\n

The system sits between incoming traffic and the network endpoints that deliver it to subscribers. It applies a layered validation process to every SMS message and voice call as it enters the network. When a communication passes validation, it reaches the subscriber\u2019s device without delay.  When the system identifies a communication as a potential threat, it alerts the subscriber. While this approach improves awareness, experience has shown that some users may still intentionally or unintentionally interact with malicious messages and fall victim to scams. To provide stronger protection, the solution has evolved to proactively block SMS identified as scams before they reach the subscriber. We adopted this approach based on its demonstrated accuracy of 0 complaint<\/strong> during the beta testing phase. This approach reduces exposure to scam attempts while maintaining high detection accuracy and preserving user choice through an opt-out mechanism.<\/p>\n\n\n\n

The AI and machine learning models use continuously updated reference datasets. This allows the system adapt to new scam tactics without waiting for manual rule updates. All inference also runs within IOH\u2019s own network infrastructure, keeping data within Indonesia\u2019s borders. No message or call content leaves the network, and no human analyst is required to review flagged communications.<\/p>\n\n\n\n

In-house build and subscriber education<\/h3>\n\n\n\n

IOH led the end-to-end design, integration, and deployment of the system as part of its network security capabilities. The AI anti-scam protection service is provided free to all IOH subscribers with an active package. IOH recognised that technology alone cannot stop every threat. It therefore paired the technical system with a subscriber education programme. The campaign teaches users how different scam types work and what to do when they encounter them.<\/p>\n\n\n\n

\n\n\n\n

Impact<\/h2>\n\n\n\n

AI anti-scam protection results in the first month<\/h3>\n\n\n\n

IOH launched the anti-scam service on 7 August 2025.  Six months after its launch, the system had successfully identified more than 2 billion calls, SMS messages, and links associated with potential spam and scam activities. It also detected over 124,000 phone numbers used for scam operations across all mobile operator networks in Indonesia, helping protect IOH subscribers from potential fraud and financial losses.<\/p>\n\n\n\n

 The solution combines subscriber awareness with proactive network protection. It alerts subscribers to potential risks while blocking SMS identified as scams before they reach the subscriber. This dual-layer approach not only prevents exposure to malicious messages but also helps users recognise common scam patterns, strengthening digital awareness as an additional layer of defence alongside network-based protection. IOH draws three lessons from the deployment. First, AI can scale anti-scam protection to a level that manual processes cannot reach. Second, close collaboration with government regulators and private partners is also essential for both effectiveness and compliance. Finally, education programmes are an integral part of effective scam prevention and not just a supplement to technical protections. IOH plans to refine the detection models and align the initiative with Indonesia\u2019s national digital transformation priorities. Overall, the IOH model shows how operators can meet data sovereignty requirements while delivering AI-enabled scam protection at national scale as part of the commitment to safeguard Indonesian subscribers.<\/p>\n\n\n\n

\n\n\n\n

Additional information<\/h2>\n\n\n\n

Participating organisations: <\/strong>Indosat Ooredoo Hutchison (IOH). Collaboration with Indonesian government regulators and private partners as part of ongoing national alignment.<\/p>\n\n\n\n

<\/div>\n \n