I’m writing this before Thanksgiving, which I used to hate and now don’t — largely because I no longer spend the day arguing with family. I’ve been listening to a lot of friends and relatives dread this year’s event because of political discussions that are likely to occur, and the equally annoying comments and critiques from those who supported both candidates.
I have no doubt that both sides will use fake news, quotes taken out of context, and false facts to make their points to the annoyance of those of us who just want to have a nice big meal and then drop into a food coma for a few happy hours afterward.
Perhaps my own most memorable Thanksgiving was going over to my then steady girlfriend’s house, having her compare herself to a Playboy playmate, and making the unfortunate comment “you wish,” to the glee of her brothers, and effectively ending the relationship. My defense of “trying to fit it,” I’m afraid, fell on deaf ears.
The week before Thanksgiving, Intel presented a fix that could be applied to this problem — an artificial intelligence tool that could prevent you from saying or tweeting something stupid. Hell, it likely could make more than a few politicians look a lot smarter than they currently do.
I’ll explain, and end with my product of the week — well, book of the week — which is an ex-Amazon employee’s take on how Amazon is going to take over the retail world. A lot of us catch up on our reading over the holidays, if only to avoid some of our more outspoken relatives and friends.
Your Personal AI Guardian Angel
In a column this summer, I referenced the two models that car makers like Toyota were using to define the future of autonomous cars. One — which drivers hate — takes the steering wheel out of the car and makes you a passenger, but the other, called “guardian angel,” just keeps you from doing stupid things.
As soon as you start doing something stupid, it disconnects you from the car and keeps you safe. For instance, a drunk driver would be no different than a child in a child seat with a fake steering wheel. Drunk drivers might think they were in control, but the car would be getting them home safely instead.
Consider this same concept applied to all of your correspondence — whether comments in email or over social media. The tool could operate in one of two modes: in one, it would insert a pause when you hit send, and ask if you really want others to view your remarks, listing the implications. Another option would be for it to reword your message or post, taking out all of the bigotry, inappropriate language and false information.
Now imagine this same capability in a Siri-like application on your phone, tied to a wireless headset. It would be listening to the same thing you were hearing, and it could do one of several things:
- It could warn you that you were likely to say something you’d regret, because its sensors would read a change in your biometrics indicating you were both angry and intoxicated.
- It could suggest an appropriate response to defuse rather than escalate a dispute — for instance, “that’s interesting, did you notice there is a bug in your potatoes?”
- Depending on the setting, it actually could provide that zinger that you otherwise wouldn’t think up until hours — if not weeks — had passed, along with the suggestion you step back to avoid being slapped or slugged.
Regarding this last possibility, and by way of personal experience, when I was around 11, my then-stepmother wanted me to go to bed at 10 on a Friday, and I wanted to stay up and watch Star Trek (yes, I’m that old). She told me that “all smart people went to bed at 10 o’clock,” to which I responded “oh yes, what time are you going to bed?” Had I stepped back, my jaw would work a lot better today. I still smile about this, granted a tad lopsided.
Wrapping Up: AI as Enhancement Rather Than Replacement
Many of today’s concerns surround the idea of AI as a human replacement. In a lengthy GAO report, and a far more concise paper from my old friend and futurist Brian David Johnson, these concerns are brought forward, and they are real.
We really aren’t ready for the sudden unemployment of millions — but what if, rather than replacing people, we instead focused AI on making people better? Kind of like the old Six Million Dollar Man, we have the technology. We can make all of us smarter and better people. Granted, it might mean the end of politics as we know it — but frankly, after the last several months, I’d say good riddance to that.
AIs could provide advice that will help us avoid mistakes, respond to difficult situations in both more timely and more appropriate ways, and either help us safely avoid nasty arguments — or if we choose, help us win them. In all cases, they could help us be better people.
That is likely a far better focus than creating AIs that will put most of us out of work, which is where the industry is mostly headed. Thankfully, Intel folks like Brian David Johnson and Intel’s resident ethnologist Genevieve Bell are working to make the right things happen. Let’s pray they are successful.
If nothing else, it could keep you from avoiding Thanksgiving family events like I do. Happy holidays from an undisclosed location in the Northwest