Conversational Artificial Intelligence; Companion or Utility?

This existential question has been discussed, analyzed in topical conversation and movies. Is conversational based Ai a friendly sassy personality or a hyper efficient task driven utility who gets the job done?

The current AI industry narrative includes hilarious episodes how you want your Ai assistant to be your new BFF. Whether its tricking Siri to say funny things or testing a consumer brand’s new bot with outrageous questions and chuckling about the replies; the view always bends to how conversational virtual assistants need to be warm and friendly and human.

Apple recently announced they’re working on ridding Siri of her robotic voice by upgrading speech synthesis markup language to code natural verbal patterns to sound more friendly and human.

Yet, it doesn’t stop at a humanlike voice, we seem to want jokes and banter too! Ask Cortana who’s her daddy, and she’ll respond with a sassy, Bill Gates centric answer while Siri can offer up funny responses to questions of what the meaning of life is or if she has a boyfriend.

These conversational quips make interacting with Ai assistants fun and offers comic relief, encouraging you to look beyond their machine learning façade. At the end of the day, the purpose of an Ai assistant is to provide a utility — whether it be troubleshooting your internet, reporting fraud, or answering your customer service questions. Imbuing the Ai assistant with the ability to joke or understand what makes a New Yorker cartoon funny, does not necessarily make them more useful.

Emotional intelligence in Ai assistants should be viewed through the lens of a utility. If you have issues completing a transaction, the assistant should identify the issue and respond appropriately the next time around. This type of emotional intelligence or context is important to creating assistants that can effectively interact with users. For example: trying to understand tone — recognizing the significance of an underlying long pause or raised voice- whether its frustration or impatience and trying to execute on the task at hand. Having the ability to understand these subtle emotional cues makes task completion easier. However, companies have taken this EQ narrative even further.

In the past year, Amazon unveiled ‘speechcons’ in Alexa, allowing the device to respond to colloquialism like beep beep. It makes for good marketing and perhaps makes the device less intimidating. However, does that make Alexa more useful in executing tasks?

Beyond functionality, there are other considerations when using this technology. As you build a still uneasy rapport with your Ai Assistant; remember this emerging category of Virtual Assistants have almost unlimited access to your data — emails, shopping habits, phone history and even online searches. As you utilize them to schedule meetings or purchase clothes, lets not forget, there’s a company behind the machine potentially listening and carefully analyzing what you say and do in the kitchen.

The hype around AI continue to center around assistants being just as intelligent as humans, we often lose sight of the fact that there’s still a lot of “artificial” in artificial intelligence. Behind every Ai assistant is a human that initially trains, programs and inputs data so that the AI engine and Machine learning algorithms learn how to respond with the goal of ultimately requiring less human intervention.

Despite situations like the Microsoft Tay mishap or IBM Watson falling flat on expectations in the healthcare sector, we have a long way to go before AI can function without human oversight. While we grapple with images of AI ranging from Her to Siri or a sassy customer service bot, its important to avoid mismatched expectations on what AI can and cannot do.

At the end of the day, lets not forget Ai based Virtual Assistants are a utility, and if you need a friend, a dog may be a better bet.

This opinion piece was originally published on The Next Web/