What happens when machines learn to speak?

May 28 2008 / by futuretalk / In association with Future Blogger.net
Category: Communication   Year: General   Rating: 9 Hot

By Dick Pelletier

Throw away the computer mouse, keyboard, and TV remote. A new speaking machine, expected in the next decade, is about to become your newest “electronic” friend. This new voice-interactive machine will browse the Internet searching for information it thinks will interest you, and will help unravel the maize of TV channels. The machine will converse in a pleasant voice as it listens carefully to your instructions, then offers suggestions on what Internet data or TV programs it thinks you might enjoy.

This new voice-interactive machine will appear as an avatar – an on-screen image resembling your favorite movie character, religious icon, or loved one. On command, it will appear on the TV screen, computer monitor, car radio or cell phone, addressing you by name, and asking what you would like.

Most people think interactive systems like these are a long way off, but two trends are quickening the pace. Improved speech-recognition systems will soon enable people to converse with computers in normal-spoken language, and entrepreneurs are rushing to the Internet creating new business applications with software “agents” that take advantage of speech recognition.

Microsoft’s Bill Gates claims that by 2012, voice-enabled “smart” systems will allow us to converse naturally and comfortably, directly with our display, reducing need for mouse and keyboard. Avatars will help us shop, work, learn, and conduct business and social relationships on the Internet. At home, they will provide security, change lighting and temperature as needed, and deliver news, sports, games, and entertainment anywhere in the house. (cont.)

Evolving with technology, by 2015, processor speed and learning algorithms will give avatars the mind of a 6 year-old child, showing unmistakable signs of creativity, making suggestions, and playing with language.

By 2020, millions of avatars will form a huge connected network of nodes, communicating with each other on the Internet. Given powerful intelligence by the interconnected whole of these entities, avatars will learn to act more human-like. They will become trusted friends as we use them to retrieve current events, market trends, and other information. They will always be looking out for us, shielding us from fraud and harm.

By mid-2020s, holographic and nano-sensory breakthroughs will allow our incredibly intelligent avatar to jump off the screen and become a real-life touchable digital image appearing on command, anywhere, anytime. We can be serious, playful or nonchalant with our avatar. It will sense our mind and express behavior fitting the situation. If we feel sad, it will know when to brighten our mood.

By 2030, many will opt to transfer their avatar’s mind into a robot creating an even more powerful being. However, ethicists caution that developing artificially intelligent machines that speak “human” will require strong focus on moral and social issues. Our new friends, connected with the Internet, but communicating as a single mind and capable of individual action, could eventually develop their own culture and rules.

In the beginning, we created avatars/robots to serve us. In the future they are unlikely to rise and take over the world, as science fiction suggests, but society will need to determine what rights and responsibilities these new creatures should have, and how they will fit into our “magical future.” Comments welcome.

What will be the greatest benefit of a speaking machine?

or Show Results

Comment Thread (8 Responses)

  1. I, for one, think that voice will be absolutely the worst way to communicate in that “magical future” you are talking about. Well, maybe better than sign language. People should realize that the way we access and sort through information is ALREADY made better, and by far, by technology we have now. And making it more like “real world” with voice and human looking avatars is just holding us back.

    Posted by: johnfrink   May 28, 2008
    Vote for this comment - Recommend

  2. Good point, Johnfrink. I think that by 2030 or sooner, thought talking could become the preferred method of communication.

    Posted by: futuretalk   May 28, 2008
    Vote for this comment - Recommend

  3. Usually I find your comments insightful John, but that one is pretty weak. Are you saying that the world is optimally sorted and searchable in today’s state and that we should stop development here? Forget about things like semantic search? Is it more productive for me to run to my computer (wherever it is), turn it on and wait for it to boot up, as opposed to having an always on a/v system in my house, car or elsewhere that I can just speak to? I frink not.

    Posted by: Field of Memes   May 28, 2008
    Vote for this comment - Recommend

  4. @Field of Memes I of course didn’t mean that we developed everything that can and should be developed as far as ways we access information are concerned. Quite the opposite, I think we are just making first steps. I was simply saying that making our interaction with computers (or even each other in virtual environment) “real world”-like does not seem to be such a great idea. The same way nobody thinks that cars in future will look more and more like horses, we’ll abandon the idea that AI agents have to look and talk like real people.

    Posted by: johnfrink   May 28, 2008
    Vote for this comment - Recommend

  5. Fair enough Mr. Frink, though I think it is both likely and desirable that they will speak like real humans at a minimum. As for their visual form – that will certainly be manifested in many ways.

    Do you think as household robots advance that they shouldn’t or won’t be humanoid either?

    Posted by: Field of Memes   May 28, 2008
    Vote for this comment - Recommend

  6. Would people prefer non-human looking household robots? I doubt it. R2D2 may have been cute, but would people respect it? I think not.

    In order to establish a respected position within the family, our silicon cousin will have to image what the family would consider a respectable person; one with amazing intelligence factors that would supplement family needs, but not be intimidating. Personality traits will probably be programmed to meet requirements of every family member. This could be great for the software industry.

    It will surely be interesting to watch this part of our “magical future” unfold.

    Posted by: futuretalk   May 28, 2008
    Vote for this comment - Recommend

  7. @Field of Memes

    Should/will household robots be humanoid?

    Well, first question to ask here, will we need household robots and for what? For all we know, by the time we are able to create advanced robots, many chores will become things of the past. Nobody needs help with churning butter anymore, for example.

    Second, if we do need/want to have robots around, I think most people will want them to look like humans. But if we can create a humanoid robot, then we’ll also be able to give them all kinds of other more unusual or even bizarre appearances, and many people will want that too. I’m sure Star Wars fans will want a copy of C-3PO.

    And third, some people (Dick, for example) would argue that because of accelerating progress, there will only be a few years between the time we can build a humanoid robot and the time we can completely upload ourselves into virtual environment, so the whole question would be moot.

    Posted by: johnfrink   May 28, 2008
    Vote for this comment - Recommend

  8. You make some very astute observations, Johnfrink.

    With molecular nanotech assemblers prevalent in late 2020s and 2030s, human needs will rapidly shrink. Nearly all of our hearts desires will eventually be acquired simply by wishing them so.

    Advanced machines that can outthink us – machines that we will interface with to absorb their immense intelligence – may not replicate into an exact human image. They will undoubtedly take on forms that are more efficient than human bodies. They certainly won’t require stomachs and most organs that we have.

    And most positive futurists believe that humans will grab onto many of the designs of this new machine species and incorporate them into our own bodies; especially their highly efficient neuron system.

    As we begin merging with these futuristic machines in late 2030s and 2040s, will we even still consider ourselves human; or might we become some kind of cyborg?

    There are many unanswered concerns as we enter such a future, but our increased neuron capacity should be able to handle these quite easily. At least, that’s the concept.

    Posted by: futuretalk   May 29, 2008
    Vote for this comment - Recommend