# AKINOLA OLATUNJI

- **Event:** [Meta Global AI Developer Hackathon](https://cerebralvalley.ai/e/meta-ai-developer-hackathon)
- **When:** Oct 26, 12:00 PM – Nov 4, 9:00 PM (EST)
- **Where:** Online
- **Team:** [Akinola Olatunji](https://cerebralvalley.ai/u/Voiceid)
- **GitHub:** https://github.com/WeeglongVoiceID/voiceid-sentinel
- **Demo video:** https://youtu.be/89V57QHcKf4
- **Gallery:** https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery
- **Page:** https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/849

# VoiceID Sentinel™ — Voice Biometric Authentication, Transaction Authorization & AI-Powered Fraud Investigation

**VoiceID Sentinel™**, developed by Weeglong AI, combines our broader voice biometric security technology with Meta AI-powered investigative intelligence to address identity-related fraud, unauthorized account access, and suspicious digital transactions.

**The Problem**

Banks, fintechs, insurance companies, and other organizations continue to face account takeovers, stolen-device fraud, compromised credentials, identity impersonation, and unauthorized transactions.

Traditional authentication methods such as passwords, PINs, and OTPs can be compromised. Even when suspicious activity is detected, security teams often struggle to investigate fragmented authentication records and determine which incidents require urgent attention.

**Our Solution**

Our underlying **Weeglong VoiceID™ platform** uses voice biometrics to verify a person's enrolled voice identity and supports a two-stage security approach:

1. **Voice Biometric Authentication:** Verify the authorized user's voice during account access, providing an additional security barrier even when a device or login credentials have been compromised.
2. **Voice-Based Transaction Authorization:** Apply additional voice verification and spoken authorization checks before sensitive transactions proceed, supported by liveness and replay-detection mechanisms.

Building on this security foundation, **VoiceID Sentinel™ introduces an AI-assisted investigative layer powered by Meta Muse Spark**.

For the hackathon, we developed a working investigation prototype that analyzes synthetic authentication-event metadata, identifies replay indicators and shared-device-token correlations, prioritizes suspicious events, and generates evidence-referenced investigative recommendations using Meta AI.

The prototype includes investigation timelines, human-review workflows, audit records, and SHA-256 evidence-consistency checks. It keeps recorded facts separate from AI recommendations, ensuring that investigators retain control over decisions.

The hackathon prototype does not itself perform biometric matching or authorize real transactions; those functions belong to the separately developed Weeglong VoiceID platform.

**Our Vision**

To help financial institutions, businesses, and government agencies strengthen identity verification, protect sensitive transactions, accelerate security investigations, and reduce fraud exposure through the combination of **voice biometric security, responsible AI, and human oversight**.

## More from Meta Global AI Developer Hackathon

- [CivicLens AI](https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/846)
- [Remont](https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/847)
- [Aketud](https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/848)
- [Andrea Conklin](https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/850)
- [Chiheb](https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/851)
- [BoardPilot](https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/852)

---

Markdown version of https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/849. Site index for agents: https://cerebralvalley.ai/llms.txt · full text: https://cerebralvalley.ai/llms-full.txt