Tech
AI can Track Unsafe Food by Surveying Customer Reviews
The researchers from Boston University developed a new AI, called BERT, that detects unsafe food. This AI uses Amazon customer reviews to tell which food is safe and which isn’t. Its accuracy is about 74% which isn’t too bad to begin with.
It detects any food that’s contaminated with chemicals, toxins, pathogens. After eating unsafe food, many of the people become ill, this AI prevents that from happening. Taking into consideration the global health program, the researchers arrived at this AI Program that uses world’s largest online retailer to help detect unsafe food items.
It follows an elaborate process. The developers of the AI linked the Amazon.com food reviews with the AI BERT(Bidirectional Encode Representation from Transformations). The 1,297,156 reviews from Amazon.com linked with the food products made it easy for the users.
Dr. Elaine Nsoesie, the assistant professor of global health at BUSPH informed that- “Health departments in the US are already using data from twitter, yelp and Google for monitoring foodborne diseases.”
The unsafe food products are a global food threat. And BERT is an intelligent way to deal with it. The WHO reports that about 600 million people suffered from illness because of contaminated food. BERT helps to track the unsafe food products and help generate food recalls.
Now the developers want to take the AI up a notch by adding not just reviews from Amazon, but product reviews from other websites and social media as well, like Twitter, Facebook. This way any product labelled foul or rotten , or anything bad gets the unsafe tag.
It isn’t the first time researchers used social media platforms for reviews. Many health organisations monitor food products by referring to the social media as well. Surveillance of food products is important and BERT is in the game to do play its role. The researchers said that BERT will stop any foodborne disease outbreaks and prevent illness and deaths.
Tech
CypherFace Targets Payment Fraud with Pre-Transaction Biometric Verification
A U.S.-based fintech company has deployed a facial biometric system that verifies user identity before digital payments are processed. CypherFace, which began commercial operations in 2024, is positioning its technology as a proactive defense against payment fraud that now costs billions annually.
Founder Syed Samir Hassan said the company developed the platform in response to the limitations of fraud detection systems that identify problems after transactions have already occurred. “Traditional fraud tools are reactive by design. They analyze patterns and flag suspicious activity, but the money has often already moved. We’re stopping it before the transaction completes,” Hassan said.
The Fraud Problem
Digital payment fraud has grown substantially despite existing security measures. Payment fraud in the European Economic Area increased to €4.2 billion in 2024, up 17% from 2023, according to data from the European Central Bank and European Banking Authority. Credit transfer fraud alone saw a 24% increase.
Synthetic identity fraud, which involves creating fictitious identities using combinations of real and fabricated personal information, has become particularly problematic. False identity cases increased 60% in 2024 compared to the previous year. These synthetic identities often pass initial verification checks because they use legitimate data elements.
Hassan said CypherFace was designed specifically to address this threat vector. “Synthetic identities work because they look clean on paper. They pass KYC checks. They build credit histories. But they can’t pass a live biometric verification tied to a real person. That’s the fundamental flaw we exploit.“
The company reports that fraudsters increasingly use AI-generated documents and deepfake technology to bypass security systems. CypherFace’s liveness detection technology is designed to identify these sophisticated spoofing attempts during the authentication process.
How the Technology Works
CypherFace provides businesses with an API that integrates into payment infrastructure. When a user initiates a transaction, the system prompts for facial verification. The technology captures and encrypts a facial scan, then applies AI-driven liveness detection to confirm a physically present individual is authorizing the payment.
The system processes the verification in real time without storing raw biometric data. Facial scans are converted into encrypted, non-reversible hashes. The platform returns only a verification result to the merchant, indicating whether the transaction should proceed.
“We designed this to be invisible to legitimate users and impossible for fraudsters,” Hassan said. “A real customer takes two seconds to verify. A criminal using a stolen card or synthetic identity can’t get past the liveness check. The math is simple.“
Deployment and Results
An e-commerce payment processor deployed CypherFace across its checkout infrastructure in late 2024. The processor was experiencing elevated chargeback rates driven by card-not-present fraud. Within 45 days of implementation, CypherFace flagged more than 1,200 fraudulent transactions that had previously bypassed existing security layers.
The integration reduced chargebacks by 62% in the monitored segment. The processor reported improved merchant satisfaction as legitimate transactions experienced minimal additional friction. The company has since expanded CypherFace to additional merchant accounts.
Hassan noted that the technology addresses a specific gap in payment security. “Most fraud prevention happens at the network level or through transaction monitoring. We’re adding a layer that asks a simple question: is the person trying to make this payment actually who they claim to be? If they’re not, the payment doesn’t happen.“
Market Expansion
CypherFace currently operates in the United States and is preparing to expand into Canada and Mexico in 2026. The company is targeting payment processors, merchant acquirers, and platforms with high transaction volumes and elevated fraud exposure.
Hassan said the company sees demand from businesses struggling with the cost of chargebacks and fraud losses. “Every fraudulent transaction costs more than the transaction value when you factor in chargeback fees, lost merchandise, and reputational damage. Businesses are looking for solutions that actually prevent fraud rather than just detect it after the fact.“
The fintech sector has broadly adopted biometric authentication, with major banks and digital financial platforms using facial recognition and fingerprint scanning for account access and transaction authorization. CypherFace is focusing specifically on payment verification rather than account login.
“We’re not trying to replace existing security. We’re adding a verification layer at the most critical point in the transaction flow,” Hassan said. “When money is about to move, we make sure the right person is authorizing it. Everything else is secondary to that.“
-
Tech5 years agoEffuel Reviews (2021) – Effuel ECO OBD2 Saves Fuel, and Reduce Gas Cost? Effuel Customer Reviews
-
Tech6 years agoBosch Power Tools India Launches ‘Cordless Matlab Bosch’ Campaign to Demonstrate the Power of Cordless
-
Lifestyle7 years agoCatholic Cases App brings Church’s Moral Teachings to Androids and iPhones
-
Lifestyle5 years agoEast Side Hype x Billionaire Boys Club. Hottest New Streetwear Releases in Utah.
-
Tech7 years agoCloud Buyers & Investors to Profit in the Future
-
Lifestyle5 years agoThe Midas of Cosmetic Dermatology: Dr. Simon Ourian
-
Health7 years agoCBDistillery Review: Is it a scam?
-
Entertainment7 years agoAvengers Endgame now Available on 123Movies for Download & Streaming for Free
