>
JUSTICE IS COMING: Ed Martin Announces Indictment Incoming On Adam Schiff!
Another Federal War on Free Speech
Study: Democrats Losing Registration Race in All 30 States that Track Party Affiliation...
Trading Up: Our Case for Trade Schools
Chinese Scientists Produce 'Impossible' Steel to Line Nuclear Fusion Reactors in Major Break
1,000 miles: EV range world record demolished ... by a pickup truck
Fermented Stevia Extract Kills Pancreatic Cancer Cells In Lab Tests
3D printing set to slash nuclear plant build times & costs
You can design the wheels for NASA's next moon vehicle with the 'Rock and Roll Challenge
'Robot skin' beats human reflexes, transforms grip with fabric-powered touch
World's first nuclear fusion plant being built in US to power Microsoft data centers
The mitochondria are more than just the "powerhouse of the cell" – they initiate immune...
Historic Aviation Engine Advance to Unlock Hypersonic Mach 10 Planes
OpenAI CEO Sam Altman Pitches Eyeball-Scanning World ID to Bankers
In the wake of a Hong Kong fraud case that saw an employee transfer US$25 million in funds to five bank accounts after a virtual meeting with what turned out to be audio-video deepfakes of senior management, the biometrics and digital identity world is on high alert, and the threats are growing more sophisticated by the day.
A blog post by Chenta Lee, chief architect of threat intelligence at IBM Security, breaks down how researchers from IBM X-Force successfully intercepted and covertly hijacked a live conversation by using LLM to understand the conversation and manipulate it for malicious purposes – without the speakers knowing it was happening.
"Alarmingly," writes Lee, "it was fairly easy to construct this highly intrusive capability, creating a significant concern about its use by an attacker driven by monetary incentives and limited to no lawful boundary."
Hack used a mix of AI technologies and a focus on keywords
By combining large language models (LLM), speech-to-text, text-to-speech and voice cloning tactics, X-Force was able to dynamically modify the context and content of a live phone conversation. The method eschewed the use of generative AI to create a whole fake voice and focused instead on replacing keywords in context – for example, masking a spoken real bank account number with an AI-generated one. Tactics can be deployed through a number of vectors, such as malware or compromised VOIP services. A three second audio sample is enough to create a convincing voice clone, and the LLM takes care of parsing and semantics.