September 2025 Summaries

AI-driven threats, particularly deepfakes, are becoming increasingly prevalent and pose significant risks to government agencies. To combat these threats, Resemble AI has introduced its Generative AI Voice-Based Deepfake Simulation Platform, now accessible to the public sector through a partnership with Carahsoft Technology Corp. This collaboration enables federal, state, and local agencies to proactively defend against deepfake attacks by providing them with advanced tools and training. The platform facilitates live phishing scenarios using hyper-realistic voice cloning and adaptive AI, allowing employees to be tested and scored on their responses to simulated threats. Carahsoft’s role as Resemble AI’s Master Government Aggregator® ensures broad access to this platform through various contract vehicles, allowing agencies to enhance their defenses quickly and efficiently.

Chatterbox, an open-source text-to-speech model, has made significant strides in the voice AI industry by outperforming existing solutions and gaining widespread popularity with over 1 million downloads on Hugging Face and 11,000 GitHub stars. Responding to community demand for broader language support, the team introduced Chatterbox Multilingual, which supports 23 languages, allowing brands, educators, and developers to create authentic and expressive voice applications beyond English. Designed for various applications, including customer support and global gaming, the model offers fine-tuned emotional expression and enterprise-level reliability, with additional features such as custom fine-tuning and real-time streaming available through Chatterbox Multilingual Pro. This development represents a major advancement in making multilingual voice AI accessible and reliable for diverse global audiences.