Artificial intelligence: Does the machine know you better than you know yourself?
Our reality is hyperconnected. Just by sliding our fingers on the screens of our phones or with just a click on the computer and we can buy, interact, share content… in essence, just being an extension of ourselves in the virtual world. In this construction of our digital self, artificial intelligence (AI) plays an important role. Especially when it comes to identity identification processes.
What is Artificial Intelligence?
The scientist, entrepreneur and director of Engineering in Google, Raymond Kurzweil gave his definition of Artificial Intelligence. He said it is “the art of developing machines with the capacity to assume functions that usually requires intelligence when humans do them”.
But this science that designs machines and systems that seem intelligent to human eyes, is not a specific phenomenon from the hyperconnected world. It actually is not even from after the Word Wide Web.
The first approach to this subject dates back to the 50’s in the twentieth century. Back then, the English mathematician Alan Turing published an article called: “Computing machinery and intelligence”. In this material, the theoretic and logical alluded the future creation of machines that could emulate the functions of human beings, especially those related to thinking actions, among them we can find the game of chess.
Afterwards, from that very same text would arise what is well-known in the world of math as the Turing test. This is nothing more than a test that evaluates the capacity of intelligent responses similar to the human ones of a machine.
Identity verification and AI
Identifying who we are on systems or automatized artefacts is possible thanks to the development of AI. Such common things as unlocking our phones with our fingerprint, authenticating with facial recognition or scanning QR codes of identification documents are possible due to the development of intelligent automatized systems.
There is a list of AI procedures for the verification of identity. Here we show you the most common and used ones all over the world:
- Documents verification: Register in databases documents such as driver’s license, that allow the search and identification of the users. This could work by the complete or partial digitalization of the document, where data is not the only element that influences, holograms or QR codes do too.
- Digital fingerprint sensors: This is one of the most recognized and used biometric systems globally. Not just because of its use but also because of its level of security, since fingerprints are unique.
- Facial recognition: This technology allows us to identify people by the biometric characteristics of their face.
- Iris recognition: It consists in the identification based on a database of people’s iris’s characteristics. It’s usually done with cameras that have infrared lights. This procedure shouldn’t be mistaken with the retina scan.
- Automatic voice recognition: It habilitates the digital devices’ response through the voice.
Challenges of the Identity verification with AI
Being just machines and codes, AI systems also have limitations. They depend quite a lot on the data used to develop them. Researchers call this process “machine learning”. Its essence is precisely that the predictive capacity of the machine is linked to the quality and precision of the data used in its programming.
“Nowadays, the best version of verification systems is the one that combines AI with expert human revision. While AI handles the most frequent and common cases (which are the majority), human experts handle the verification of those cases that identify that AI is less trustworthy or simply cannot operate (these usually are the most atypical cases)”.Rubén Francisco Manrique, in “Artificial Intelligence and its impact in identity verification”
Experts in this field refer that the main challenge of AI in relation to identity verification is related with security. The frauds based on identity thefts are a very sensitive matter, as of today.
Another recurring concern is the agility and scalability of these new systems. Processes must be more and more agile and must occur in comfortable platforms for the users, all these in balance with reasonable consumption of computational resources.
Does the intelligence of softwares that “read” your face or predict your behavior amaze you? Subscribe to this blog where there is more information that can help you build your digital self.