Tech weekly: No laughing matter this — Amazon, other cos are betting big on voice recognition

The machines are here and they’re freaking people out- that seemed to be a big theme this week with Amazon’s intelligent personal assistant Alexa laughing randomly and scaring users across the US, raking up the debate on how much intelligent devices are invading our lives.

Amazon’s smart device Echo is widely used in the US and lets users make it perform tasks like making phone calls, answer queries and controlling home devices using voice commands. The assistant Alexa replies, much like Apple’s Siri or Samsung’s S Voice.

“Intelligent” machines and devices, especially those that reply and perform tasks based on voice commands, are among the biggest bets for consumer-based technology companies.

Everyone from Google to Microsoft to Amazon and Apple are investing in working around natural human language and understanding accents to be able to reach the widest possible user base. The market is pegged at over USD 18 billion in the next five years.

While the commercial interest and opportunity are justified, voice-based recognition is a Pandora’s box of issues and the Alexa problem is probably indicative of the larger issues that lurk around this.

Amazon said the random laughter came from the voice assistant recognising some phrases as “laugh” and the company is now working on making Alexa laugh only when asked “Alexa can you laugh?” so as to create fewer false positives. The machine will also reply “yes I can” before it actually laughs.

Google, arguably the most popular search engine, has said previously that over 20 percent of all searches are now voice based. It is also rolling out voice search in regional languages in India, making it easier for first time users who may not be able to read or write, use its search features better.

The convenience and commercial case is compelling for corporates and with better machine learning, systems are becoming more and more intelligent and taking human-like decisions.

This widespread use raises issues of who owns the user voices being processed, how much control users have over the sound commands being analysed, the possibility of voice-based frauds, companies collecting regular voice conversations (like the famous Samsung Smart TV issue), and how it feeds into the larger engine for machines to “learn” how the human brain works.

Automobiles also have systems that work on voice commands and there is again little or no clarity on who owns the voice history being recorded by the vehicular systems.

These are interesting times for technology and intelligent machines are making our lives simpler and easier. However, there is a compelling case to be made for being a little more aware of how all these voice-based systems are using data from our personal lives and invading our privacy.

The government or regulatory bodies will always take time to catch up with technology, and it is about time we as users become more aware and begin to question the usage of the data that we are generating.