Showing posts with label deepfake. Show all posts
Showing posts with label deepfake. Show all posts

Artificial Intelligence - What Is The State Of Biometric Security And Privacy?


Biometrics is a phrase derived from the Greek roots bio (life) and metrikos (measurement).

It is used to examine data in the biological sciences using statistical or mathematical techniques.

In recent years, the phrase has been used in a more precise, high-tech sense to refer to the science of identifying people based on biological or behavioral features, as well as the artificial intelligence technologies that are employed to do so.

For ages, scientists have been measuring human physical characteristics or behaviors in order to identify them afterwards.

The first documented application of biometrics may be found in the works of Portuguese historian Joao de Barros (1496–1570).

De Barros reported how Chinese merchants stamped and recorded children's hands and footprints with ink.

Biometric methods were first used in criminal justice settings in the late nineteenth century.

Alphonse Bertillon (1853–1914), a police clerk in Paris, started gathering bodily measurements (head circumference, finger length, etc.) of prisoners in jail to keep track of repeat criminals, particularly those who used aliases or altered features of their appearance to prevent detection.

Bertillonage was the name given to his system.

After the 1890s, when it became clear that many people had identical dimensions, it went out of favor.

Richard Edward Henry (1850–1931), of Scotland Yard, created a significantly more successful biometric technique based on fingerprinting in 1901.

On the tips of people's fingers and thumbs, he measured and categorized loops, whorls, and arches, as well as subcategories of these components.

Fingerprinting is still one of the most often utilized biometric identifiers by law enforcement authorities across the globe.

Fingerprinting systems are expanding in tandem with networking technology, using vast national and international databases as well as computer matching.

In the 1960s and 1970s, the Federal Bureau of Investigation collaborated with the National Bureau of Standards to automate fingerprint identification.

This included scanning existing paper fingerprint cards and creating minutiae feature extraction algorithms and automatic classifiers for comparing electronic fingerprint data.

Because of the high expense of electronic storage, the scanned pictures of fingerprints, as well as the categorization data and minutiae, were not kept in digital form.

In 1980, the FBI made the M40 fingerprint matching technology operational.

In 1999, the Integrated Automated Fingerprint Identification System (IAFIS) became live.

In 2014, the FBI's Next Generation Identification system, an outgrowth of IAFIS, was used to record palm print, iris, and face identification.

While biometric technology is often seen as a way to boost security at the price of privacy, it may also be utilized to assist retain privacy in specific cases.

Many sorts of health-care employees in hospitals need access to a shared database of patient information.

The Health Insurance Portability and Accountability Act emphasizes the need of preventing unauthorized individuals from accessing this sensitive data (HIPAA).

For example, the Mayo Clinic in Florida was a pioneer in biometric access to medical records.

In 1997, the clinic started utilizing digital fingerprinting to limit access to patient information.

Today, voice analysis, face or iris recognition, hand geometry, keystroke dynamics, gait, DNA, and even body odor combine with big data and artificial intelligence recognition software to rap idly identify or authenticate individuals based on voice analysis, face or iris recognition, hand geometry, keystroke dynamics, gait, DNA, and even body odor.

The reliability of DNA fingerprinting has evolved to the point that it is widely recognized by courts.

Even in the absence of further evidence, criminals have been convicted based on DNA findings, while falsely incarcerated prisoners have been exonerated.

While biometrics is frequently employed by law enforcement agencies, courts, and other government agencies, it has also come under fire from the public for infringing on individual privacy rights.

Biometric artificial intelligence software research has risen in tandem with actual and perceived criminal and terrorist concerns at universities, government agencies, and commercial enterprises.

National Bank United used technology developed by biometric experts Visionics and Keyware Technologies to install iris recognition identification systems on three ATMs in Texas as an experiment in 1999.

At Super Bowl XXXV in Tampa, Florida, Visage Corporation presented the FaceFINDER System, an automatic face recognition device.

As fans entered the stadium, the technology scanned their faces and matched them to a database of 1,700 known criminals and terrorists.

Officials claimed to have identified a limited number of offenders, but there have been no big arrests or convictions as a result of such identifications.

At the time, the indiscriminate use of automatic face recognition sparked a lot of debate.

The Snooper Bowl was even dubbed after the game.

Following the terrorist events of September 11, 2001, a public policy discussion in the United States focused on the adoption of biometric technology for airport security.

Following 9/11, polls revealed that Americans were prepared to give up significant portions of their privacy in exchange for increased security.

Biometric technology were already widely used in other nations, such as the Netherlands.

The Privium program for passenger iris scan verification has been in effect at Schiphol Airport since 2001.

In 2015, the Transportation Security Administration (TSA) of the United States started testing biometric techniques for identification verification.

In 2019, Delta Air Lines, in collaboration with US Customs and Border Protection, provided customers at Atlanta's Maynard Jackson International Terminal the option of face recognition boarding.

Passengers can get their boarding cards, self-check baggage bags, and navigate TSA checkpoints and gates without interruption thanks to the technology.

Only 2% of travelers choose to opt out during the first launch.

Biometric authentication systems are currently being used by financial institutions in routine commercial transactions.

They are already widely used to secure personal smart phone access.

As smart home gadgets linked to the internet need support for safe financial transactions, intelligent security will become increasingly more vital.

Opinions on biometrics often shift in response to changing circumstances and settings.

People who support the use of face recognition technology at airports to make air travel safer may be opposed to digital fingerprinting at their bank.

Some individuals believe that private companies' use of biometric technology dehumanizes them, treating them as goods rather than persons and following them in real time.

Community policing is often recognized as an effective technique to create connections between law enforcement personnel and the communities they police at the local level.

However, other opponents argue that biometric monitoring shifts the emphasis away from community formation and toward governmental socio-technical control.

The importance of context, on the other hand, cannot be overstated.

Biometrics in the workplace may be seen as a leveler, since it subjects white-collar employees to the same level of scrutiny as blue-collar workers.

For usage in cloud security systems, researchers are starting to build video analytics AI software and smart sensors.

In real-time monitoring of workplaces, public spaces, and residences, these systems can detect known persons, items, sounds, and movements.

They may also be programmed to warn users when they are in the presence of strangers.

Artificial intelligence algorithms that were once used to create biometric systems are now being utilized to thwart them.

GANs, for example, are generative adversarial networks that replicate human users of network technology and applications.

GANs have been used to build fictitious people's faces using biometric training data.

GANs are often made up of a creator system that creates each new picture and a critic system that iteratively compares the fake face to the original photograph.

In 2020, the firm Icons8 claimed that it could make a million phony headshots in a single day using just seventy human models.

The firm distributes stock images of the headshots made using its proprietary StyleGAN technology.

A university, a dating app, and a human resources agency have all been clients.

Rosebud AI distributes GAN-generated photographs to online shopping sites and small companies who can't afford to pay pricey models and photographers.

Deepfake technology has been used to perpetrate hoaxes and misrepresentations, make fake news clips, and conduct financial fraud.

It uses machine learning algorithms to create convincing but counterfeit videos.

Facebook profiles with deepfake profile photographs have been used to boost political campaigns on social media.

Deepfake hacking is possible on smartphones with face recognition locks.

Deepfake technology may also be used for good.

Such technology has been utilized in films to make performers seem younger in flashbacks or other similar scenarios.

Digital technology was also employed in films like Rogue One: A Star Wars Story (2016) to incorporate the late Peter Cushing (1913–1994), who portrayed the same role from the original 1977 Star Wars picture.

Face-swapping is available to recreational users via a number of software apps.

Users may submit a selfie and adjust their hair and facial expression with FaceApp.

In addition, the computer may mimic the aging of a person's features.

Zao is a deepfake program that takes a single picture and replaces the faces of stars from movies and television shows in hundreds of video.

Deepfake algorithms are now being used to identify the deepfakes' own videos.

~ Jai Krishna Ponnappan

You may also want to read more about Artificial Intelligence here.

See also: 

Biometric Technology.

Further Reading

Goodfellow, Ian J., Jean Pouget-Abadie, Mehdi Mirza, Bing Xu, David Warde-Farley, 

Sherjil Ozair, Aaron Courville, and Yoshua Bengio. 2014. “Generative Adversarial Nets.” NIPS ’14: Proceedings of the 27th International Conference on Neural Information Processing Systems 2 (December): 2672–80.

Hopkins, Richard. 1999. “An Introduction to Biometrics and Large-Scale Civilian Identification.” International Review of Law, Computers & Technology 13, no. 3: 337–63.

Jain, Anil K., Ruud Bolle, and Sharath Pankanti. 1999. Biometrics: Personal Identification in Networked Society. Boston: Kluwer Academic Publishers.

Januškevič, Svetlana N., Patrick S.-P. Wang, Marina L. Gavrilova, Sargur N. Srihari, and Mark S. Nixon. 2007. Image Pattern Recognition: Synthesis and Analysis in Biometrics. Singapore: World Scientific.

Nanavati, Samir, Michael Thieme, and Raj Nanavati. 2002. Biometrics: Identity Verification in a Networked World. New York: Wiley.

Reichert, Ramón, Mathias Fuchs, Pablo Abend, Annika Richterich, and Karin Wenz, eds. 2018. Rethinking AI: Neural Networks, Biometrics and the New Artificial Intelligence. Bielefeld, Germany: Transcript-Verlag.

Woodward, John D., Jr., Nicholas M. Orlans, and Peter T. Higgins. 2001. Biometrics: Identity Assurance in the Information Age. New York: McGraw-Hill.

What Is Artificial General Intelligence?

Artificial General Intelligence (AGI) is defined as the software representation of generalized human cognitive capacities that enables the ...