That we live in a digital-first age is no myth – with smartphones, social media, digital money, and a host of other technologies playing a ubiquitous part of our lives. But companies such as Apple, Google, Microsoft, Samsung, and some other start-ups would like us to believe that the ‘Artificial Intelligence or AI-first’ age is here with machine learning technology taking center stage. What these companies have in common are talking and ‘thinking’ AI-powered digital assistants that take sci-fi fantasy to a whole different level. Their value proposition? To increase your day-to-day personal and professional efficiency by quickening the actions you already take on your smartphone and learning from your behavior to make proactive recommendations.
But have we reached the stage of true artificial intelligence or are these digital assistants no good than marketing gimmicks that fail to deliver on their promise? Let’s see how they compare to human virtual assistants.
#1 AI needs TONS of data to get its facts right and THIS limits its potential Understanding concepts, recognizing behavior patterns, and interpreting the non black-and-white signals doesn’t come naturally to AI – given its very DNA. It needs not just more, but FAR MORE data than humans to get its learning right and this forms the basis of its limitations. Apple’s Siri, Google’s voice search assistant, Microsoft’s Cortana, and the likes from other tech biggies are often good at their job because these firms have tons of data in the form of consumer voice searches, behavior patterns, etc. to feed into their machine learning tools and create useful solutions. The problem arises in cases where digitized data is sparse – for instance, mental health issues or situations where users express personal crisis. In a recent study conducted on Siri, Google Now, and S voice digital assistants, all of them failed to recognize sentences relating to domestic abuse, violence, suicidal thoughts, and were unable to direct users to suicide prevention/mental health helpline numbers or take other remedial action.
#2 Though designed to enhance productivity, multi-tasking isn’t exactly a digital assistant’s strong suite
Apple’s Siri lacks an open API structure making it completely incompatible with other apps – meaning you can’t track your projects on a platform like Podio, open a podcast in your playlist, or do sundry other things while giving directions to Siri. Google Now has limited capacity to interoperate with some third-party apps while Amazon’s Alexa is one of the minority popular digital assistants built on an open API – making it easy for users to tie in any service or app into it. But it lacks its own smartphone platform and is only all about Amazon – would be good for you if all you want from life is to be a receiver of Amazon’s services. Doesn’t quite cut it, right? Human virtual assistants on the other hand are adept at multitasking – document/email management, online research, travel and lifestyle management, corporate administration – you name it!
#3 Digital assistants force you to talk the way they want, not how you do naturally
It’s embarrassing to talk to digital assistants and that’s why only 6% of people who use their services are willing to do so in public – 51% prefer to activate these ‘helpers’ behind closed doors. The barrier of ‘talking to technology’ is accentuated as most digital assistants force users to talk the way they understand as voice recognition technology still has its limitations when it comes to understanding dialects, tones, and phrases that are not commonly used. What’s worse – if you have more than one digital assistant – one that controls smart devices in your home, another that sets reminders and alerts, and yet another that controls let’s say, your entertainment options (TV, music, etc.) – you are likely to find yourself in a comedy of errors kind of situation. Not only do these different digital assistants fail to interact with each other and various smart home appliances, they often fail to understand a user’s natural use of language. And then, there’s the privacy issue with AI assistants – they fail to keep your conversations private, at least for now.
The Bottom-Line: For digital assistants to become truly smart, unobtrusive, and proactive helpers to make life easier, it will take a revolution – one that is aimed at overhauling their skills and eliminating blond spots – not at utilizing the treasure trove of cheap and available data to mask the inaccuracies. As for human personal assistants, comparing them to their digital counterparts borders on complete ignorance of the work and responsibilities these integral members perform for an organization or individual.