InvestSMART

Speak and you shall find

There are many machines you can talk to today, but none of them will answer convincingly enough to fool you into thinking they are human. Fortunately, most of them are still useful.
comments Comments
Upsell Banner

You don't have to go far to get the best business education degree these days. In my case, it is only 1.8 miles - or so Siri, my "humble personal assistant", advised when I asked her the other day where the best such learning could be found.

"Education" was about the only word Apple's new voice-recognition software on my iPhone 4S seemed able to recognise. Guessing I was looking for somewhere nearby with a high rating from reviewers, she came up with a list of educational establishments, headed by Dominican University (four stars), close to my home in California.

A manual search revealed that Dominican does offer an interesting "Green MBA" - though it is not the top US business school ­according to the instant response to a typed Google search. That took me less time than asking Siri a question (I asked several while trying to elicit a sensible answer), making me feel the feature touted by Apple as the coolest on its latest smartphone was a bit of a gimmick at this stage.

Which is a shame. It would be nice to think we are closer to machines displaying enough artificial intelligence to pass the Turing test. Set by mathematician Alan Turing in 1950, it requires machines to answer questions convincingly enough to fool someone they were human.

The nearest we have come so far is IBM's Watson computer understanding natural-language questions and beating two human contestants this year in the US quiz show Jeopardy. A commercial product is expected from this research project in the next two years.

Meanwhile, the S***ThatSiriSays website continues to add the ready-made answers discovered by users. For example, saying "I need to hide a body" prompts the response: "What kind of place are you looking for? Reservoirs, mines, dumps, swamps ..."

Apple has put a beta label on Siri, which is rare for a company that likes to release mature technologies, but it is an admission that there are holes in Siri's knowledge. It cannot give sport scores, flight information or much data about places outside the US, for example.

It is good at telling you the weather or time anywhere, converting currencies and measurements, finding local restaurants and services, and giving directions. It can carry out commands on the iPhone, such as playing a song, setting alarms, recording notes and reminders, and replying to messages.

In those ways it does seem a friendly personal assistant, and we are beginning to treat our gadgets more like real people. We touch them, talk to them or gesture at them - and get a response. Merely looking at the latest Samsung Galaxy Nexus smartphone will make its Android 4.0 operating system unlock the phone using facial recognition.

Siri can take dictation wherever you might use a keyboard on the iPhone 4S, but its results were indifferent and required corrections as it misheard words. There are other options for the 4S and other models, but currently Siri features only on this iPhone.

I like Dragon Dictation from Nuance. Although Siri uses ­Nuance's technology as well, this free app offers comprehension in more than 20 languages, whereas Siri can handle only English, French and German for now. It is available for the iPad, iPod touch and Android devices, where it is known as FlexT9.

Dragon Go is a voice-powered search app that understands the context of your queries and links you to the most relevant services in a series of tabs. For example, saying "Tickets for the San Francisco Giants", the baseball team, brings up the Stubhub sports service.

Google also allows voice in its search app and for searching on Google TV, while Microsoft's Kinect controller for the Xbox also brings voice commands to the television. Google Translate lets you speak and hear a translation - acting almost as an interpreter on the phone - and works in 60 languages.

Voice control is being encouraged in the car to keep the driver's hands on the wheel. Ford, the motor company, has installed its Microsoft-powered Sync voice-recognition system to power ­music and other services in even its cheapest models. Bluetooth headsets from the likes of Jawbone and Plantronics allow spoken commands and offer text-to-speech services.

Plantronics' new Voyager, Marque and M50 headsets feature its Vocalyst services, allowing listening to news feeds, checking the weather, email or text messages, and posting to Facebook and Twitter, all hands-free.

In November, Pioneer announced Zypr, a voice-control platform that will span car, home, office and smartphone.

I am also impressed by other technologies that can listen intelligently. Shazam and Soundhound identify artists and songs by listening to a few seconds. Lyrics, and an option to buy the music or tickets, can be included in the results.

Similarly, IntoNow, acquired by Yahoo and now available as an iPad app, listens to TV shows you are watching and brings up relevant material, such as movie reviews and cast lists.

This example may be both voice and face recognition, but I like how I can now go into a local coffee shop and, instead of producing a wallet or my phone, simply say "Put it on Chris's tab" to buy a latte.

This is down to Square's Card Case app. The mobile payment service allows me to open a virtual tab with a merchant, showing my photo and name on their terminal as my phone comes in range of the cafe. Saying my name merely confirms to them my identity, and my registered card is charged.

The hard human-machine interface is finally melting away, like the milky wisp on top of my coffee.

Music to your ears: apps that identify a tune, enhance TV viewing and help you shop

SoundHound
There's little between Shazam and SoundHound in their ability to identify the music you hear playing, but SoundHound also allows "query by humming" for that tune in your head. An attractive iPad app for this music identification service, as well as Android and iPhone ones, a link with Spotify to allow users to get straight to listening to the music that's found, and LiveLyrics, which follows the song line-by-line, all hit the right notes.

IntoNow
This TV-watching app was bought by Yahoo in April and launched this month on the iPad - a natural home now that tablets are being used as remote controls and constant companions in the living room. It listens to the sound from programmes and brings up relevant information. You can also broadcast what you're watching through the app, see what friends are viewing and chat with them about what's on TV.

Square Card Case
Square, the mobile payment service, allows users to store their accounts with merchants in this virtual credit card case. Using location data on a smartphone, it can tell a retailer when you are approaching the store, allowing them to greet you, serve your favourite drink and let you pay by just saying your name. It also allows you to view receipts, browse menus and use built-in mapping to discover new places.

Copyright The Financial Times Limited 2011.

Share this article and show your support
Free Membership
Free Membership
Chris Nuttall - Financial Times
Chris Nuttall - Financial Times
Keep on reading more articles from Chris Nuttall - Financial Times. See more articles
Join the conversation
Join the conversation...
There are comments posted so far. Join the conversation, please login or Sign up.