All news with #deepfake detection tag

👓 Smart glasses are returning with advanced sensors and AI, creating new privacy and security challenges for users and bystanders. They can record or livestream covertly and feed footage to AI systems for face recognition and data retrieval, enabling stalking, fraud, and surveillance. Platform policies and outsourced review raise additional exposure. Mitigations include updates, permissions control, MFA, and disabling AI training where possible.

🔎 Deepfake technology is now widely accessible and sufficiently realistic to fool employees, executives and automated heuristics. A 2025 Gartner survey found nearly half of cybersecurity leaders encountered audio or video deepfakes in the prior year, and real-world incidents show attackers using synthetic media for both financial fraud and reputational sabotage. Organizations must combine rapid forensic verification, coordinated legal action and clear communications while pursuing long-term authentication and watermarking standards to restore trust.

🛡️ As deepfakes and injection attacks evolve, identity verification must move from isolated media checks to end-to-end session trust. Ricardo Amper of Incode explains that high-fidelity synthetic faces, replayed footage, virtual cameras, rooted devices, and automated probing can all defeat perception-only defenses. Incode Deepsight combines perception, integrity, and behavioral signals in real time to validate the entire verification session and reduce false acceptances while blocking persistent unauthorized access attempts.

🔍 This article explains how modern deepfakes are created, deployed, and detected in real-world scams, and why virtually anyone can be a target. It describes common visual, auditory, and behavioral signs—lighting and lip-sync errors, unnatural blinking, electronic vocal tones, and awkward gestures—and notes attackers use tools from Telegram bots to commercial services like HeyGen and ElevenLabs. Practical advice includes ending suspicious chats, verifying identities via alternate channels, agreeing a family codeword, tightening privacy on photos and recordings, enabling strong account security, and using content-analyzer services to flag AI-generated media.

🛡️ The World Economic Forum warns that advances in deepfake and face‑swapping technologies are enabling attackers to bypass KYC and remote verification, creating financial and systemic risks. A WEF Cybercrime Atlas study examined numerous face‑swap and camera injection tools and found that low‑latency, high‑fidelity real‑time swaps can be delivered into verification pipelines. While many tools were designed for creative use, researchers found some capabilities that defeat traditional KYC protections, though detectable artefacts like temporal desynchronization, lighting and compression inconsistencies provide practical detection targets. The report issues 27 recommendations and urges providers, fraud teams and regulators to evolve defences in step with generative AI.

⚠️ The FBI is warning of AI-assisted fake kidnapping scams that use fabricated images, video, and audio to extort victims. Criminal actors typically send texts claiming a loved one has been abducted and follow with multimedia that appears genuine but often contains subtle inaccuracies. Examples include missing tattoos, incorrect body proportions, and other mismatches, and attackers may use time-limited messages to pressure victims. Observers note the technique is currently of uncertain effectiveness but likely to be automated and scaled as AI tools improve.

🔍 The UK Information Commissioner’s Office (ICO) has asked the Home Office for urgent clarity after a National Physical Laboratory (NPL) report identified racial bias in the retrospective facial recognition (RFR) algorithm Cognitec FaceVACS-DBScan ID v5.5 used by police. The study found far higher false positive rates for Asian (4%) and Black (5.5%) subjects compared with white subjects (0.04%), with an observed disparity between black males (0.4%) and black females (9.9%). Deputy information commissioner Emily Keaney said the ICO was disappointed it had not been informed earlier and stressed that public confidence, transparency and proper oversight are essential while the Home Office moves to operationally test a replacement algorithm.

🔍 A new report from Entrust analyzing over one billion identity verifications between September 2024 and September 2025 warns that fraudsters increasingly use generative AI to produce hyper-realistic digital forgeries. Physical counterfeits still account for 47% of attempts, but digital forgeries now represent 35%, while deepfakes comprise 20% of biometric frauds. The report also highlights a 40% annual rise in injection attacks that feed fake images directly into verification systems.

🔍 AI can now produce highly convincing receipts that reproduce paper texture, detailed itemization, and forged signatures, making manual review unreliable. Expense platforms and employers are deploying AI-driven detectors that analyze image metadata and transactional patterns to flag likely fakes. Simple countermeasures—users photographing or screenshotting generated images to remove provenance data—undermine those checks, so vendors also examine contextual signals like repeated server names, timing anomalies, and broader travel details, fueling an ongoing security arms race.

📸 Google has added support for the C2PA Content Credentials standard to the Pixel Camera and Google Photos apps on the new Pixel 10, enabling tamper-evident provenance metadata for images, video, and audio. The Pixel Camera app achieved Assurance Level 2 in the C2PA Conformance Program, the highest mobile rating currently defined. Google says a combination of the Tensor G5, Titan M2 and Android hardware-backed features provides on-device signing keys, anonymous attestation, unique per-image certificates, and an offline time-stamping authority so provenance is verifiable, privacy-preserving, and usable even when the device is offline.

📸 Google is integrating C2PA Content Credentials into the Pixel 10 camera and Google Photos to help users distinguish authentic, unaltered images from AI-generated or edited media. Every JPEG captured on Pixel 10 will automatically include signed provenance metadata, and Google Photos will attach updated credentials when images are edited so a verifiable edit history is preserved. The system works offline and relies on on-device cryptography (Titan M2, Android StrongBox, Android Key Attestation), one-time keys, and trusted timestamps to provide tamper-resistant provenance while protecting user privacy.

📷 Google announced Pixel 10 phones will embed C2PA Content Credentials in every photo captured by the native Pixel Camera and display verification in Google Photos. The Pixel Camera app achieved Assurance Level 2 by combining Tensor G5, the certified Titan M2 security chip, and Android hardware-backed attestation. A privacy-first model uses anonymous enrollment, a strict no-logging policy, and a one-time certificate-per-image strategy to prevent linking. Pixel 10 also supports an on-device trusted timestamping mechanism so credentials remain verifiable offline.

📚 The online world often mirrors the schoolyard, and bullying can intensify when a new term begins. A 2023 Microsoft study highlights cyberbullying as a top parental concern, with harassment ranging from name‑calling and rumor‑spreading to sextortion and deepfake images. Watch for behavioral changes, keep open, nonjudgmental lines of communication, and review app privacy settings. If abuse occurs, calmly teach children to block, capture evidence and report incidents to platforms and schools.