Google We Have A Problem With Duplex: Cool tech, but needs bounding

Have you seen the really cool demo of Google Duplex? This is a fantastic demonstration that combines voice recognition, machine learning and very capable voice synthesis to show how AI can interact with people to get things done. Watch the video and you can see for yourself how very smooth this AI bot can be:

The use cases here are very widespread. It makes my imagination really go wild. I would love an AI assistant I could set loose to get things done for me. Imagine being able to task your AI to do things like:

  • Call to set up a doctor appointment or dentist appointment
  • Set up meetings with your staff
  • Book reservations for restaurants, movies, hotels
  • Wait on line for a human to answer at the insurance company or bank or government agency, saving your hours and hours
  • Call your politicians to give them an ear full of what you think
  • Call everyone who signed up for an event to confirm they are really coming
  • Call to tell the service desk at your car dealer that the car is making a strange noise and they need to come and fix it
  • Call the airlines to complain about the poor service
  • Call the taxi company to schedule a cab to come
  • Finding and calling a plumber or electrician or any other provider
  • Call every company that might know something about me and tell them I want my information removed from their databases in accordance with the GDPR

Really the list is endless.

But my fear is people on the receiving end of those calls will not always be pleased to learn that an AI is calling them. In fact, if people start calling me with their AI I might actually lose my sense of humor over it. What if people and businesses start calling me with their AI to do things like:

  • Telemarketing of anything
  • Pretending to be my neighbor or pretending to be someone else I would have an affinity with
  • Inviting me to events or to support causes
  • Asking me to donate to politicians, maybe even using the voice of the politician and acting like I”m really talking to my Congressman.
  • Any interaction with a fraud or a conman

If you are in a services business, I think you may want to take calls from AI, it might mean more business. But some businesses might not want these calls. And any of us should have an ability to decide if it is ok to interact with us using these robots, don’t you think?

One thing for sure, if a computer is deceiving you to intentionally make you think it is an AI, that is not good. Do you agree?

We have always advocated that we technologists play a role in helping others think through the ethics of the technology world. This new Google feature just underscores how important this is.

We have also advocated that we all stretch our imagination by enjoying SciFi every once in a while. I’m thinking I might have seen one alternative future of how this may turn out (see Westworld).

There is other SciFi that also comes to mind. Remember Blade Runner? The Voight-Kampff test? Maybe in the very near future every call you get should start out with a series of questions for the caller to assess their reaction and determine their humanity.

“The tortoise lays on its back, its belly baking in the hot sun, beating its legs, trying to turn itself over but it can’t. Not without your help. But you’re not helping.” “What do you mean I’m not helping” “I mean you are not helping…”

For more on AI and related topics see:


, , ,

Leave a Reply