Link
“We could basically replicate any person.”
That's an attention-getting claim. The article does not consider the security problems associated with this technology. But this is the first thing that springs to my mind. So much of our social fabric is built on the implicit security mechanisms of visual, auditory and other forms of sensory identification. Imagine a hacker cloning your spouse with this technology and then performing a man-in-the-middle attack through Skype. "Honey, could you give me the bank account # again, I seem to have lost it." All such attacks seem far-fetched until they happen in real-life and then people realize in hindsight that of course such an attack would occur once the cost of the equipment and software was less than the amount of money which could be stolen using that equipment.
Of course, the broader importance of this technology is in reducing impedance in the human-computer interface (HCI). The human visage, voice and body language already are the ideal human interface. As we continue along the explosion of raw computational power for the foreseeable future (with or without Moore's law), the surplus computation that you don't need to check your email, browse the web or play media (all activities requiring very little computation) is going to be invested into improving the HCI. We are just about to cross the threshold where more people will be talking to their devices than touching them. Technology like this demo'd by Soul Machines is going to become pivotal in continuing to mold computing machines to human needs instead of the other way around.
Update: AI Creates Fake Obama
No comments:
Post a Comment