News
Imposter Voices are Making Their Way into Our Lives
Researchers use fluid dynamics to detect artificial imposter voices in deepfake audio.
Consider the following scenario: A phone call comes in. When an office worker answers the phone, he hears his boss, who is in a panic, tells him that she forgot to transfer money to the new contractor before leaving for the day and needs him to do it. She gives him the wire transfer information, and the crisis is averted with the money transferred.
The employee reclines in his chair, takes a deep breath, and watches as his boss enters the room. The person on the other end of the phone line was not his boss. It wasn’t even a human being. He was hearing the voice of an audio deepfake, a machine-generated audio sample designed to sound exactly like his boss.
Such attacks using recorded audio have already occurred, and conversational audio deepfakes may not be far behind.
Deepfakes, both audio and video, have only recently become possible due to the advancement of sophisticated machine-learning technologies. Deepfakes have added a new layer of uncertainty to the digital media landscape. Many researchers have turned to analyzing visual artifacts – minute glitches and inconsistencies – found in video deepfakes to detect deepfakes.
This is not Morgan Freeman, but how would you know if you weren’t told?
Audio deepfakes may pose an even greater threat because people frequently communicate verbally without using video, such as through phone calls, radio broadcasts, and voice recordings. These voice-only communications greatly expand attackers’ ability to use deepfakes.
To detect audio deepfakes, we and our University of Florida colleagues devised a technique that compares the acoustic and fluid dynamic differences between voice samples generated organically by human speakers and those generated synthetically by computers.
Natural vs. synthetic voices
Humans vocalize by forcing air through the vocal tract’s various structures, which include the vocal folds, tongue, and lips. By rearranging these structures, you can change the acoustical properties of your vocal tract and produce over 200 distinct sounds, or phonemes. However, the acoustic behavior of these different phonemes is fundamentally limited by human anatomy, resulting in a relatively small range of correct sounds for each.
Audio deepfakes, on the other hand, are created by first allowing a computer to listen to audio recordings of a specific victim speaker. Depending on the exact techniques used, the computer may only need to hear 10 to 20 seconds of audio. This audio is used to extract important information about the victim’s voice.
The attacker chooses a phrase for the deepfake to speak and then generates an audio sample that sounds like the victim saying the chosen phrase using a modified text-to-speech algorithm. This process of creating a single deepfaked audio sample takes only a few seconds, potentially giving attackers enough flexibility to use the deepfake voice in a conversation.
Deepfake audio detection
Understanding how to acoustically model the vocal tract is the first step in distinguishing human speech from deepfake speech. Scientists, thankfully, have techniques for estimating what someone – or some being, such as a dinosaur – would sound like based on anatomical measurements of its vocal tract.
We did the opposite. We were able to extract an approximation of a speaker’s vocal tract during a segment of speech by inverting many of these same techniques. This effectively allowed us to peer into the anatomy of the speaker who created the audio sample.
From there, we hypothesized that deepfake audio samples would be free of the same anatomical constraints that humans face. In other words, deepfaked audio samples were analyzed to simulate vocal tract shapes that do not exist in humans.
Our test results not only confirmed our hypothesis but also revealed something new. When we extracted vocal tract estimations from deepfake audio, we discovered that they were frequently comically incorrect. Deepfake audio, for example, frequently produced vocal tracts with the same relative diameter and consistency as a drinking straw, as opposed to human vocal tracts, which are much wider and more variable in shape.
This realization shows that, even when convincing to human listeners, deepfake audio is far from indistinguishable from human-generated speech. It is possible to determine whether the audio was generated by a person or a computer by estimating the anatomy responsible for creating the observed speech.
Why is this important?
The digital exchange of media and information defines today’s world. Everything from news to entertainment to conversations with loved ones is usually done digitally. Deepfake video and audio, even in their infancy, undermine people’s trust in these exchanges, effectively limiting their usefulness.
Effective and secure techniques for determining the source of an audio sample are critical if the digital world is to remain a critical source of information in people’s lives.
Source Credit: Conversation.com
On May 24th, Brian May, a founding member of the rock band Queen who later earned a Ph.D. in astrophysics, posted several images to Instagram of a crop circle seen near Marlborough, England.
“Have you noticed anything out of the ordinary here in the English countryside?” The photos were captioned by May. “I’d never seen a crop circle before. As a result, I’m always skeptical of them. But yesterday, as we flew back from our production rehearsal space, over a location near Marlborough, there was this. […] Who creates these fascinating works of mathematical art? Is it a hoax? Are they created by extraterrestrials? And… how…? And what is their goal?”
Responses to May’s post have been mixed, with some claiming that the phenomenon is paranormal, while others believe that hoaxers are to blame.
Crop circles have sparked speculation in the modern era since at least the mid-1970s, with theories ranging from hoaxers to otherworldly beings to “earth energies.”
Despite the fact that people have claimed responsibility for certain crop circle formations, mysterious circles of flattened plants discovered in fields date back much further than modern-day hoaxers.
W.Y. Evans-Wentz recorded folktales of faeries coming in the night to thresh farmers’ grain in his 1911 book The Fairy Faith in Celtic Countries. Similarly, in 1678, an English woodcut pamphlet depicts ‘The Mowing-Devil,’ who is shown mowing crops in a circular pattern.
While some dismiss these as folkloric inspiration for modern-day hoaxers, others see them as proof of a phenomenon that predates man-made imitation.
No one has yet claimed responsibility for the crop circle photographed by May.
Source Credit: SingularFortean.com
According to a press release issued by the Department of Defense (DoD):
Due to the passage of the National Defense Authorization Act for fiscal 2022, Deputy Secretary of Defense Kathleen Hicks, in consultation with the Director of National Intelligence (DNI), amended her original directive to the Under Secretary of Defense for Intelligence & Security on July 15, 2022, by renaming and expanding the scope of the Airborne Object Identification and Management Group (AOIMSG) to the All-domain Anomaly Resolution Office (AARO).
USD(I&S) Hon. Ronald S. Moultrie informed the department today of the establishment of AARO within the Office of the Under Secretary of Defense for Intelligence and Security, and named Dr. Sean M. Kirkpatrick, most recently the chief scientist at the Defense Intelligence Agency’s Missile and Space Intelligence Center, as its director.
The AARO’s mission will be to coordinate efforts across the Department of Defense and other federal departments and agencies in the United States to detect, identify, and attribute objects of interest in, on, or near military installations, operating areas, training areas, special use airspace, and other areas of interest, and, as needed, to mitigate any associated threats to operational safety and national security. Anomaly, unidentified space, airborne, submerged, and transmedium objects are included.
Under Secretary of Defense for Intelligence and Security (USD(I&S)) Ronald Moultrie will lead the AARO Executive Council (AAROEXEC), which will provide oversight and direction to the AARO along the following primary lines of effort:
1. Surveillance, Collection and Reporting
2. System Capabilities and Design
3. Intelligence Operations and Analysis
4. Mitigation and Defeat
5. Governance
6. Science and Technology
This newly reported expansion of the Pentagon’s UFO investigation program follows low congressional trust in their investigative efforts.
Following the release of the much-anticipated preliminary assessment report on UFOs by the Office of the Director of National Intelligence last year, many in the intelligence community were critical of what they saw as the report’s failure to offer any concrete explanations for most of the incidents examined, particularly in light of concerns about secret Russian or Chinese technology.
The Pentagon then promised to revamp the task force in charge of investigating UFOs, which resulted in the formation of the AOIMSG, which has since been renamed the AARO.
This reflects Congress’ growing interest in UFOs, which was most recently demonstrated during a House Intelligence Subcommittee hearing on the subject last May—the first of its kind in more than 50 years.
The congressional hearing allowed lawmakers to question the Pentagon about unidentified aerial phenomena (UAP)—the current government term for UFOs—and for government officials to explain their current position and outline plans to investigate the issue further.
During the hearing, there were few mentions of extraterrestrials, though the Pentagon did express a particular interest in reports containing unusual flight characteristics such as incredible speed, transmedium capabilities, and undetectable means of propulsion.
Since the existence of the Pentagon’s Advanced Aviation Threat Identification Program (AATIP), which reportedly ran from 2007 to 2012, was made public in 2017, congressional interest in UFOs has skyrocketed.
Interest in the encounters between Navy pilots and UFOs grew, and in 2019, several senators, including Sen. Mark Warner (D-Va.), then vice chairman of the Senate Intelligence Committee, were briefed on them.
The Senate Intelligence Committee, led at the time by Senator Marco Rubio (R-Fla.), then included a directive in their Intelligence Authorization Act for Fiscal Year 2021 ordering the Director of National Intelligence (DNI) to create a report on “unidentified aerial phenomena” in consultation with the Secretary of Defense.
That bill resulted in the formation of the Unidentified Aerial Phenomena Task Force (UAPTF), which was in charge of producing the aforementioned preliminary assessment report.
Source Credit: SingularFortean.com
Alien Encounter
‘Queen’ Founding Member Shares Crop Circle Picture
DoD Announces Expanded Effort to Investigate UFOs
‘Queen’ Founding Member Shares Crop Circle Picture
Cryptid Community Intrigued by Recent Photo of ‘Ogopogo’
Woman Spots Alien in Backseat of Her Car
-
News7 months ago‘Queen’ Founding Member Shares Crop Circle Picture
-
News7 months agoCryptid Community Intrigued by Recent Photo of ‘Ogopogo’
-
News7 months agoWoman Spots Alien in Backseat of Her Car
-
News8 months agoCamera Captures Massive Predator In Georgia Swamp
-
News7 months agoAustralian Man Spots Dogman on Fishing Trip
-
Uncategorized7 months ago“Bedroom Abduction” Caught on Camera
-
News7 months agoMan Claims ‘Black Hole’ in Spine Predicts End of World
-
News7 months agoThe Marlborough Monkey is a Cryptid Fans Classic