As time goes on, technology gets more and more advanced. It seems that, with every year, new devices get even closer to being the perfect creation of mankind. In fact, for years now, the possibility of true artificial intelligence has never seemed closer.
Despite that fact, it still has not been created, however, many creations have come amazingly close. Siri, as one example, is Apple co’s designated voice for almost all of its devices. Despite the surprising cleverness of Siri’s answers and remarks, she still lacks some basic elements of true AI including independent thought.
Progress towards even smarter devices does not stop there. Now, in most devices, programs exist that can recognize both your face and your voice. With every device, these programs become smarter and smarter, but how smart are they now? Despite all this, there are consequent loopholes and interesting deficiencies in these programs.
Facial recognition often works by the program’s ability to analyze and remember the most prominent facial “landmarks” of the user.
Voice recognition works differently. As you speak, you create vibrations in the air, or rather, analog- waves. These waves will then be collected and translated by an ADC.
Does your tone of voice affect a program’s ability to understand? Would a fake recreation of someone’s face ( a picture or mask) be recognized as the real person? Of course, how else would one know the answers for sure without testing?
An interesting way to test the voice recognition programs would be to say the same thing twice; once with a higher-pitched tone and another with a low-pitched tone. Other tones can be used as well. Possibly even a song-like tone.
Another test that could provide valuable information is using a picture of a person for his own facial recognition program. It does not have to be limited to pictures however, closely related family members with the same defining facial features can also be used to test the program’s intelligence.