arXiv:2606.24523v1 Announce Type: cross Abstract: Scam phone calls exploit vulnerable communities worldwide, yet research on detection has focused almost exclusively on English and other high-resource languages. In low-resource settings such as Turkish, detection is especially difficult, as annotated data is scarce and technological defenses remain limited. This research investigates how large language models (LLMs) can support scam detection in Turkish by introducing the first public multi-modal dataset of 100 aligned audio-transcript pairs of scam and benign conversations. We evaluate seven
The proliferation of AI-powered scam techniques and increasing global connectivity necessitate more robust and inclusive defense mechanisms against fraud, particularly in non-English contexts.
This development offers a potential pathway to enhance scam detection capabilities in previously underserved languages, leveraging LLMs to protect vulnerable populations and financial systems on a global scale.
The availability of the first multi-modal dataset for Turkish scam detection and the exploration of LLMs for this purpose marks a step towards more equitable and effective cybersecurity measures in low-resource language environments.
- · Turkish-speaking consumers
- · Financial institutions
- · AI cybersecurity firms
- · NLP researchers
- · Scammers
- · Cybercrime networks
Improved detection of phone call scams in Turkish leads to fewer successful scam attempts and reduced financial losses for individuals.
The methodology and dataset inspire similar efforts in other low-resource languages, fostering a more global and inclusive approach to AI-driven fraud detection.
The enhanced security measures could inadvertently lead scammers to develop more sophisticated, multi-channel attack vectors, prompting an arms race in AI-powered defense and offense.
This signal links to a primary source. Continuum Brief monitors and indexes it as part of the live intelligence stream — we do not republish source content.
Read at arXiv cs.AI