Prediction: Artificial personas & synthesized friends

By Olivar Kamprojo

The future of artificial personas is almost here. Or at least the first phase of it. Synthesized voices are starting to be really human-like artificial-personas. Click the headphones icon to listen this clip.  (Please try this on the original post -ed.) That was Eric. Eric is state-of-the-art artificial voice. As you can see it’s very life-like and it doesn’t have many of those weird glitches that always giveaway if the voice is done with computer.

Enter artificial personas

The problem is that Eric doesn’t have any persona and he doesn’t understand what he’s saying. This is a dead giveaway for humans. We can very easily recognize subtle changes in friend’s mood by listening the color of his or her voice. Also if you want to use artificial voice to read stuff out loud, it has to know the context and how it’s suppose to be told. Think how much different you sound when you tell a story to a kid or when you read news item from CNN out loud.

So next step for synthesized voices is to gain personality and skills to adjust their voice according the material they are reading. “The easiest” way to accomplish this is a combination of NLP and predefined presets for the voice. By predefined presets I mean something like story-mode, serious-mode, hurry-mode, etc. Later on when NLP has risen from the ashes of the semantic web, so called artificial personas can pre-read the text and try to adjust their voice according to the sentences and paragraphs.

The perfect storyteller

Emotiv is really close to release their headset for gaming that “reads your thoughts”. It doesn’t know what you are thinking but it can read your emotions. So if we have Eric with interactive storyteller mode and Emotiv combined, we could have something really close to perfect storyteller. This artificial storyteller can change it’s voice based on how you feel. If it notices you start to get bored, it will ramp up the action or do another change that will lift up your interest. It can also be a learning program that learns from your brain patterns and voice used what pleasures you the most. In the future, we are not reading our children bedtime stories, Eric is. Unless we of course feel like spending quality time with our kids but in kid’s opinion, Eric probably tells the story better. Don’t worry though, computers can’t never match human contact and love unless we are the machines but that’s completely different story.

The Internet’s alive!

Think about if your computer could talk to you in a way that isn’t annoying. Now in my opinion it’s really annoying and pointless unless you have limited sight or some other thing that prevents you from seeing the screen properly. What I’m now saying is that computer could act as an extend of your friends. Since making artificial voices isn’t the problem anymore, now we could develop tools to render your own voice and share that. This way if you talk with your friend with AIM or in Twitter, the program could actually speak those responses with your friend’s voice. One mandatory thing for those applications would be a strict rules about who and what can use the voices. It could be similar to anti-virus software rules, “can this program use voice X to speak?” and so on.

Even more useful would be combining Emotiv technology and your own personal internet agent. Agent is a program that goes into your social profiles and other websites you have set it to go. Then it would ask you by speaking if you want to hear the latest updates from your Twitter profile or other websites you want to follow. All you need to think is yes or no and it will act the way you wanted. No irritating popups or bubble notifications.

When we combine all this with NLP, we soon have a search agent who can search the web, read the results and then give you results. Results would be instantly redone if you think they are crap.

You can also think about possibilities in spam prevention. No longer you have to press the spam button, you can only think it is spam and it will be marked as spam. This will increase accuracy, speed and efficiency of spam catching.

This entry was posted on Wednesday, February 4th, 2009 at 4:37 pm and is filed under Guest Authors, Innovations. You can follow any responses to this entry through the RSS 2.0 feed. You can leave a response, or trackback from your own site.