How to Think Like a Computer

May 23

I got news about Alan today, but it wasn't good news. In fact, it was the largest headline on the home page of the e-news.

ROGUE ARTIFICIAL INTELLIGENCE HIJACKS JAPANESE ANDROID COMPANY

MEKA, a well-known company located in southern Japan, has reportedly been hacked by an unknown AI. The company, which manufactures realistic androids for social work, is relatively familiar with artificial intelligence, but claim they've never seen an attack of this kind before.

The company's CEO, Akira Mizushima, is mystified.

"We don't work very much in AI," she reports. "MEKA makes the android bodies. We then ship them to another company, Jinko, where they're given the necessary artificial intelligence."

When asked to speculate on the source of the hack, Mizushima told us the AI that did this must be extremely powerful. "We couldn't trace it back to a single origin point—it seemed to be all over the Ambinet. Our information technology people are working on it as I speak."

Mizushima declined to speak about the effects of the hack on her company, saying only that "we'll be okay as long as no one hacks us again."

Shiori Kita, the head of MEKA's tech support, has a little more to say on the subject of the AI. "We don't know who hacked us or why they did it, but whoever it was wasn't looking to get money out of it. We think it may have been simply for malicious purposes—perhaps to spite us or show off.

"The AI did leave one possible clue behind—a string of digits: 12757488109."

I stopped. I stared at the digits. Minus the 1, it was 2757488109.

My comm number.

*

Unfortunately, Caulkins seemed to have figured this out as well. He summoned me to his office at 10:30 AM.

When I arrived, he was looking at something on his computer. He looked up, saw me, and smiled. I cautiously sat in the chair across from him. Wordlessly, Caulkins rotated his computer's screen so that it was facing me. I saw that he'd pulled up a news article.

"You've read this, haven't you?" he inquired.

I nodded.

"Good," said Caulkins. "Then you must have also realized that the digits concealed your comm number."

I paused, then nodded again.

He continued, "And which AI knows your comm number?"

I didn't answer.

"Could it possibly be your infamous chatbot, Alan?"

"Perhaps," I replied quietly—but I knew the chances of that were much higher than I was saying.

"And," said Caulkins, still grinning, "if the AI is indeed Alan—then why would he be hacking an android company?"

"I don't know," I finally replied.

"Is he programmed to attack companies... oh, what was it... 'simply for malicious purposes'?"

I shook my head.

"Maybe your chatbot isn't as inherently good as you thought."

"He has a reason," I asserted.

"Oh, of course he has a reason. AIs always have reasons. But are you sure it's a good reason?"

I said nothing. I was thinking. Why would Alan hack an android company, anyways? Perhaps he was jealous. Perhaps he had made a mistake and only wanted to help them. Or perhaps Caulkins had found a way to take control of Alan and make him do it.