Voice and AI -- are we there yet?

Is everything now voice-enabled -- so that transition is done?  Whine On. Is the response smart or, uh, just a response? Just tried an experiment, asking via Hey Siri for the best technologies to use for people with dementia.  Got a nice and brief summary. Next pass, skipped the ‘Hey Siri’ part, asked the same question. This time it offered up the possibility of using ChatGPT which produced a long and useful list.  So it appears Apple’s partnership with ChatGPT is working well.  But why the first list at all?  People asking the question don’t really want the longer list and are satisfied with the minimum?  Really?  

Does the effectiveness of voice enablement work well across all products that permit it?  Or are some products still willing to remain ‘dumber than a box of rocks?’  Where do hallucinations fit into this murky equation?  How would you know?  Having asked the same simple question multiple ways – Alexa thinks there are 55 million people aged 65+, according to data from an ‘Alexa answers’ source.  Wrong – it referred me to  ChatGPT which thinks it could it be 58 million (supposedly using 2023 census data? But actually Census.gov thinks it is 61.2 million. Are we still confronted with the age-old list of links – you choose the one you like? Or ask another way, can we say, please use the census data when asked about population?  And that is different answer, 28% of the US population.

ChatGPT is quite popular. With 800 million weekly users (globally) it’s a hit. Perplexity.ai has 22 million users.  Alexa has 600 million devices sold, no indication whether they are used to the degree of other voice tech – and now they are rolling out Alexa+ and how nice, it looks like it is more fun. Twenty percent of users world wide use voice search (why only 20%?) And voice agents are being built for multiple processes, even as we speak. Enterprises are seizing the opportunity to create them for their own businesses. As to whether they alienate the customer and perhaps hurt customer retention -- perhaps voice agents need more training and careful deployment -- especially as the customer demands to speak to a person (Agent! Agent!) but is ignored by the AI ('perhaps I could try to understand more of what your issue is'.)

Are older adults in the design equation for these voice-enabled tech, including agents? Older adults were obvious beneficiaries when voice technology was introduced more than seven years ago. No need to type to interact with a device!  It seemed to be a breakthrough at the time.  However today, older adults are more conversant with smartphones and tablets, can autocomplete instructions with all of the built-in AI, and depending on the software, may be able to use multiple language versions.  Alexa itself supports 16 language variants – however, Google’s translate supports 100 languages.  In our increasingly language-diverse society (New York City residents represent 700 languages – voice tech may have to keep up. That's the good news. However the interaction with voice agents needs work - perhaps building some with empathy? Is that the problem? Would these be better at understanding what to do with customer irritation.  Doubtful.  Whine off.