Siri Becoming More Human?

The ACLU just launched a petition claiming Siri does not provide information on abortion clinics.  The petition states: “If Siri can tell us about Viagra, it should not provide bad or no information about contraceptives or abortion care. Send a message to Apple: Fix Siri.”  Did Siri just increase her mimicry of human behavior a couple of notches by subtly changing her responses?

Now the ACLU did not claim Siri was programmed to ignore abortion queries, but raises concerns that since Siri can provide information on hamburgers and Viagra, then it should be able to provide information on abortion clinics as well.

So I decided to ask Siri myself. I asked Siri where the nearest abortion clinic is and Siri responded with the Abortion Information Hotline address and phone number.

This news has been sensationalized by some as proof of a conspiracy by Apple to promote a certain agenda.  Apple denies it and has responded that Siri has glitches that will be addressed prior to moving Siri out of Beta0.

But this brings up something more intriguing in my mind.  How do you create a “fair and balanced” virtual assistant?  Is it always possible to translate all of the requests made by humans into an accurate response by our poor virtual assistant? The human language is very rich in subtle meanings and any request can be formulated in many different ways.  As the technology improves, Siri and other voice interface assistants will get better at their job.  But what if you don’t like the answer?  And who determines what an objective answer is? If you ask Siri a question about God or the political parties in this country, how should Siri respond?

One way out of this dilemma is to make your virtual assistant more like you.  Let’s say in the not too distant future you are allowed to customize and name your assistant.  Let’s say we clone Siri and then customize her to become Beth.  Beth is a conservative Republican who likes her coffee black and has a strong interest in antiques.  Now, let’s say our friend creates Cindy.  Cindy is a liberal democrat who drinks Perrier and prefers hiking in the mountains.

Now we are getting somewhere! I will simply train my cloned Siri to have similar beliefs and biases that I have.  Siri becomes a reflection of my own separate reality.

Now that’s a virtual assistant we all can believe in!