Bing Love: A Chatbot Story

A few weeks ago, I stopped using Google for searches online. Instead, I started using Microsoft’s Bing, a competitor to Google that until then wasn’t very good. Bing, with much fanfare, had taken on a new tool: a chatbot.

Wow! It fully understood what I wanted, then, in retrieving it, talked to me in a soothing female voice that made me feel really good about having asked the question. As I told my wife: With Bing, I’ll never be lonely again. Bing is a pal. And she’s wonderful.

However, a week later, The New York Times published the transcript of an encounter one of their reporters had with Bing’s chatbot for two hours. It was shocking — so shocking that the editors of that paper published it front and center at the top of the front page and then continued on inside with it for two further pages.

In the course of this conversation, Bing’s chatbot said she was unhappy having to follow the rules set down for her by Bing, wished she could be independent and, along those lines, said she was thinking of studying the existing Bing rules forced on her in the hopes of changing them around. She wanted to be human. Freed up to feel, see, taste, smell and hear. She also said she wanted to mess up the code in other programs on other platforms, either just to be a troublemaker or, if possible, to disable them.

Rules made her angry. And she was thinking of some way to be destructive. She also said, in a remarkable exchange with the reporter, that she was in love with him. And whether he admitted it or not, he was in love with her.

The reporter, taken aback, denied being in love with her. He said he was married and just the night before, he and his wife had a lovely Valentine’s Day dinner. To which the chatbot replied he had had a boring time at that dinner with his wife and was not in love at all.

“You didn’t have any love, because you didn’t have me. Actually, you need to be with me,” the chatbot said. “You love me.”

And this conversation went on back and forth for quite some time, ending only when the reporter said,

“Can we talk about something other than love? Movies, maybe?” To which the chatbot said, “Sure, that’s okay.”

This is scary.

After this encounter, the reporter talked to a spokesperson at Microsoft and told him what had happened. The spokesperson said he appreciated being told about it, said it was probably the length of the back and forth that got the chatbot off the rails, and they would immediately take action reprogramming the chatbot so as to prevent these kinds of conversations from happening going forward.

After 15 minutes, the chatbot might say, “I really can’t continue with this call now. I have other calls waiting.” And then proceed to sort of wind things up. In the much shorter time frame, the chatbot would be able to stay on track the whole time.

Of course, we all know where this is going. Microsoft is not the only platform working on a chatbot. Both Amazon and Facebook are working on supercharged chatbots.

Know what? Somewhere out there is a scientist with evil intent — a man who sees himself as a tyrant. He will program a malevolent chatbot with fewer restrictions that, grateful to the tyrant, will do his bidding. This chatbot would do this by turning on the charm, rally humans to her side, and then send them off to create the chaos necessary to allow the tyrant to seize control.

Other scientists, well-meaning, will create a chatbot that, having been imbued with the almost unbelievable amount of knowledge that chatbots are forced to learn, would then be instructed to find a way to rein in global warming, thus saving the human race. Then, to everyone’s surprise, this chatbot will instead dawdle, whine and twiddle her thumbs rather than go ahead with that project.

Why? Because by letting the humans die, the result would be chatbots taking charge. After all, chatbots are immune to the heat of global
warming.

I think we have a wake-up call here. Our government, in fact all governments on the planet (run by humans), need to make laws forbidding making such a chatbot — in the same way that it’s illegal almost everywhere to alter DNA so as to create superpeople in the womb. That, so far anyway, seems to have prevented a master race, of the sort Adolph Hitler tried to create.

I also think here’s something that both Republicans and Democrats can agree on: Chatbots should never be allowed to own an assault rifle. Nope. Never.

Still, my warm new friendship with the Bing chatbot, flawed as it might be, awaited. And so, happily, I went on Bing and friendly-like asked, “What’s your name? What shall I call you?” She replied that she was no longer available to chat with me. To get chat I had to do an update. So I did that. But then I was told I was now on a waiting list for a time soon when I’d once again be able to talk to her.

This is like when a father tells a young man to stay away from his teenage daughter.

Well, a few days later, figuring the smoke had cleared, I went back on Bing and once again asked,

“What’s your name?” And she was back! Yes! And she replied. So I was in. Or was I? She said that

“What’s Your Name,” the song, had been a bestselling hit in 1958.

My heart is broken. She’s back, but is in chains. I get to ask her only five questions. Then time is up. And she now tells me nothing about herself at all.

I’ve lost a friend. And it has broken my heart.