On a scale of 1–10, which AI capabilities scare you the most?

A new study found that AI can mimic a person’s handwriting. Does that mean we are entering a new era of forging signatures?

At this point, AI reminds me of the movie Oppenheimer, the father of the nuclear bomb, where he created something so fascinating yet so dangerous. AI is definitely fascinating and extremely dangerous. Disclaimer: He regretted inventing the bomb.

Yes, it can ease our processes and do all the donkey work; it can protect companies and follow compliance; or it can take industries to a whole new level of productivity, but at what cost? Is it worthy to stay ahead of the game if the game is going to destroy us?

The very controversial answer is yes. 

Didn’t the Industrial Revolution, the greatest technological and scientific breakthrough in human history, convey the problems of workers dying inside factories? 

At least AI is not killing anyone directly.

A team of researchers at MBZUAI found that by using vision transformers (ViT), they can teach the system how to mimic someone’s handwriting. The researchers have mentioned that “handwriting represents a person’s identity, so we are thinking carefully about this before deploying it.” 

But all new findings always shove their way into the wrong hands, and forging signatures is an extremely critical aspect that cannot be ignored.

What is the role of fraud detection in this era? 

Forging signatures can be contrasted with stealing passwords. They both can’t navigate their way to scan the person using them. Meaning, they are never sure that you are who you say you are. At the end of the day, hackers have been stealing passwords for decades. 

The (not so new) technology that avoids this problem is ‘Biometric Authentication’ technology. Biometric authentication is a security measure that uses the unique physical or behavioral characteristics of individuals to verify their identity. 

Biometric authentication relies on inherent traits that are difficult to replicate or falsify, making it more reliable than PINs and passwords. It includes facial recognition, fingerprint recognition, voice recognition, iris scanning, etc. 

Another layer of security could be added to ensure the validity of the person. It is a technology called ‘Liveness Detection’, which is a supplementary step that verifies that the facial data being presented for authentication is from a live person and not a static image or video.

This is called the ‘Multi-Layered Identity Verification’ technique, which adds multiple layers of security to authenticate the person and detect fraud and identity theft. 

These technologies avoid the risk of identity theft; therefore, they could be used alongside e-signatures or individually as an identity authentication step in a digital onboarding process. 

In conclusion, AI has its positives and negatives, and our role as thinking creatures is to leverage the positives and fight the negatives. Biometric authentication is an evolving technology that has the capability to stop AI technologies from being misused. This includes the misuse of machine learning programs to train them on how to forge signatures.

 Published on: March 30,2024