Dangerous Brains

by Erik Hamre

  Once the attacks were over, though, the artificial intelligence simply slipped away, it disappeared, vanished.

  It vanished into empty cyberspace, and it left no traces that it had ever been anywhere.

  This behaviour surprised the experts from DARPA and the government. They had expected the first human level artificial intelligence to make multiple copies of itself, and to then hide these copies in the cloud and various computer systems across the world - everything to ensure that no one could ever exterminate it. But the artificial intelligence Kevorkian had created did nothing as expected.

  Everything it did was unpredictable.

  Its behaviour was erratic and random.

  It was the most beautiful thing Vladimir had ever seen.

  “Why did he create it?” Sarah asked.

  “I was hoping you could help me answer that question,” Vladimir responded.

  Sarah Kevorkian recalled her last conversation with her ex-husband. It had only been a week earlier, on May 25th, and Kevorkian had seemed to be his usual self. They had met up for a coffee in Los Angeles, and Sarah had updated Kevorkian about how her new life was going, and how she was coping with being a mum again. She had given birth to a pair of twin girls the previous fall, after struggling to get pregnant with her new husband, an investment banker, for years. They were now divorced as well, but at last he had given her the kids she had wanted before they parted ways. Kevorkian had seemed genuinely happy for her when they met up, and she had agreed that he would get a chance to meet her new kids quite soon.

  “So there was nothing odd about his behaviour?” Vladimir asked.

  “You tell me,” Sarah replied. “Did he act any differently in the office last week?”

  Vladimir stared at the floor. She was right. Vladimir had spent countless hours with Kevorkian in the office over the last month or so, but he hadn’t noticed anything odd. Kevorkian had been Kevorkian; Brilliant.

  Impatient.

  Unpredictable.

  “No, he didn’t,” Vladimir replied.

  “I’ve got one question, though,” Sarah said.

  “Yes?”

  “Why are you here? Why is it that you are asking me these questions?”

  Vladimir shrugged his shoulders. “I don’t know. They wanted someone who knew Andrew. Someone who knew him well. I have known him for more than twelve years.”

  “You’re not an American citizen, Vladimir. You’re Russian. You have worked with Andrew on this brain project for what? The last ten years. And then it turns out he has tricked you all along. Of all the competent people in Neuralgo, why would the government pick you? Why would they tell you military secrets, secrets of national importance? Don’t you think that’s odd?”

  Vladimir hadn’t considered the dilemma before then. When Ronald Kraut had explained they needed the one person who knew Kevorkian the best, Vladimir had instantly known he was that person. He hadn’t considered him not being an American as an issue. For the last twelve years he had considered himself as American as anyone.

  “No, I don’t think so,” Vladimir replied.

  “Well, then you’re not as smart as I thought you were,” Sarah said.

  13

  1st of June 2015

  DARPA’s remote Listening Station No 3

  The Nevada Desert

  DAY 1:

  1600 Hours

  “Why did you pick me?” Vladimir asked.

  “You know why,” Kraut replied.

  “I really don’t. So please enlighten me.”

  “I know who you are, Vladimir. I know you’re an agent. I have known for some time.”

  Vladimir wasn’t sure what shocked him the most when Kraut said he believed Vladimir was an agent. The casual tone it was said in, or the fact that Kraut and the government actually believed he was an agent. Regardless, the result was that Vladimir felt his entire world come crashing down around him in an instant. Kraut believed Vladimir was an agent for the SVR, that he was a spy. How could Kraut possibly believe that? And for how long had he believed it? Or did he actually believe it at all? Maybe this whole thing was just an elaborate setup to make Vladimir incriminate himself?

  “We have you on tape,” Kraut said. “The meeting you had with Grosevski, the night before flying to the US with Kevorkian in 2003.”

  Ronald Kraut placed his mobile phone on the desk in front of Vladimir, located the audio player, and pressed the play button.

  “That recording means nothing. I just said what I had to say to get out of the country,” Vladimir explained, after having listened to the first ten seconds of the recording. He remembered the conversation by heart. Before he had flown to the US he had been approached by a government official from Kremlin. The government official had shown Vladimir a picture of Vladimir’s family, and told him to report back to Russia if one of Kevorkian’s businesses ever achieved any major technological breakthroughs.

  “Did you ever report back?” Kraut asked.

  “No. Never,” Vladimir said, his entire body shaking.

  “Not even when you spoke to your brother three months ago?”

  Vladimir’s face instantly went pale. “But that…. that was my brother. I didn’t mean to… I didn’t mean to.”

  “I know. I’ve listened to the conversation. And I believe you, Vladimir. Still, it is probably sufficient to get you convicted as a spy. Thirty years in prison. Minimum.”

  Vladimir didn’t reply. Instead he threw his arms across the chest and curled his lips. Every instinct in his body told him to keep his mouth shut, to not say another world until he got a lawyer. “What do you want from me?” he still asked, after only a few seconds of uncomfortable silence.

  “What has just transpired isn’t just going to be an American problem, Vladimir. If and when this escalates we need someone who can communicate our concerns to the Russian government, someone with credibility.”

  “You want me to be the go-between? You want me to be the one to tell the Kremlin that an American artificial intelligence has become unstoppable?”

  “If it comes to that, and it’s still only an if at this stage. But yes, we want you to inform the Russian Government that we have to shut down our defense system. We want you to be the one to explain that there is no other way, and that they should probably do the same. Hopefully we will avoid this situation. But if we don’t, you might be the only person who can stop us from exterminating ourselves in our attempt to save humanity.”

  “They will never believe me.”

  “Maybe not, but at the moment you are our best bet. So let’s just pray it will never get that far.”

  Vladimir closed his eyes. Americans: They always resorted to praying when things got tough.

  The problem was that there was quite possibly a new God in the equation. If Kevorkian had truly been successful in creating the first Artificial General Intelligence, then it wouldn’t take long before it would achieve God-like powers. And this God, this God would probably be just as cruel and unjust as the ones most people prayed to. To the first Artificial Super Intelligence, humans would be insignificant, insignificant and an annoyance. Just like cockroaches were an annoyance and a vermin in our kitchens, humans were a vermin on Earth. Up until now we’d had no natural enemies, none that were superior to us anyway, and no one who could stand up against us when we polluted and populated like there was no tomorrow.

  We were like the agent says in the Matrix movie ‘a virus,’ a virus that would eventually compete with the Artificial Super Intelligence for the Earth’s resources. And the Artificial Super Intelligence would need lots of resources. It was a machine. It would need to keep cool as it rapidly created copies and incremental improvements of itself.

  There was only one end game.

  Eventually it would have to deal with the human problem.

  It probably wouldn’t exterminate the human race.

  We didn’t exterminate other species just because we used the same resources.

  But it would
have to decimate the population.

  Make it controllable.

  Sustainable.

  And the best ways to do that: Start World War III, create a deadly flu, or simply wipe us out through some method we didn’t even have the mental capacity to understand yet.

  The only other option it had was to leave Earth and let the humans eventually exterminate themselves.

  But that option was the least likely to happen. It would be a lot easier to deal with the human problem head on.

  It wasn’t like humans uprooted everything and moved to another country if they got bogged down by a few insects.

  Humans dealt with the problem.

  And just the same, the AI would deal with us.

  “What happens to Sarah?” Vladimir asked.

  “Nothing,” Kraut replied. “She will have to stay here until the situation is resolved though. Nobody leaves these premises. No exceptions.”

  “Fair enough. And that means I will have access to her if I need it?”

  “Yes, you will. You will also have access to the rest of the team in a couple of hours. I’ve assembled a team of experts who will assist us in shutting down the AI before it reaches the next levels of intelligence.”

  “What experts? I would prefer to have some of my colleagues from Neuralgo here. They are the best in the industry on artificial intelligence.”

  “There won’t be any other AI experts on the team. You and I are it. Studies have shown that people with varied backgrounds have a higher likelihood of solving unprecedented problems. So our team, Team Cronus, will consist of people with varied backgrounds. They have all been pre-selected on the basis of Protocol Cronus.”

  “You’ve got to be kidding me. This isn’t some teambuilding exercise, Kraut. If you want us to have any realistic chance at all of shutting this thing down, then we need to have access to the best people. That means Neuralgo guys.”

  “We have already been assigned the best people – as chosen by Protocol Cronus.”

  Vladimir looked away. He knew it was a big mistake. But there was nothing he could do.

  14

  1st of June 2015

  DARPA’s remote Listening Station No 3

  The Nevada Desert

  DAY 1:

  1800 Hours

  Team Cronus had been pre-selected by the author of Protocol Cronus, Ronald Kraut, back in 2013. The list hadn’t included any names. It had just stated:

  The most prominent scientists or experts in the following fields:

  -Nanotechnology.

  -Artificial Intelligence.

  -Psychology.

  -Computer security (hacking).

  And that was it; a team of four. There had of course been more detailed instructions on how to deem who the current top experts in those fields were, and the list of prospective names had to be updated bi-annually. But the protocol left a lot open to personal interpretation. The military could be one of the most bureaucratic organisations in the world during peace time. In war time it had to be flexible though, it had to be fiercely adaptable. And so it had been decided that if Protocol Cronus was ever invoked, Team Cronus had to have the most flexible structure of any military organisation. Decisions would have to be made on the fly, and there would have to be a minimum of command chains.

  This principle was even applied when it came to the selection process of team members. If there wasn’t enough time to get the most prominent Psychologist on the team, the protocol clearly stated that Kraut could choose the next candidate on the list. It was how Vladimir had been selected.

  Vladimir had been the third choice on the Experts on Artificial Intelligence List, but he had been bumped up to number one by Kraut. The tipping point had been Vladimir’s Russian background and his executive position at Neuralgo. Kraut had never believed Vladimir was an agent, but he had always known it could become a good negotiation card if the situation ever required it.

  The two other candidates, the psychologist and the computer expert, had both been Protocol Cronus’ first choices. Kraut had worked with them for more than a year and a half already. They had both been on DARPA’s payroll as informal advisors for the last twelve months. Kraut had wanted to learn whether he could trust them if it ever became necessary to call them in. Not that he had ever expected it to happen.

  Kraut himself was the leading expert on nanotechnology. There was no one disputing that fact.

  “So you’re the author of Protocol Cronus? You’re the one who has made all these decisions?” Vladimir asked, staring at Kraut. For a second Kraut seemed to hesitate. Vladimir had been to several presentations where Kraut had been the keynote speaker. Without exception Vladimir had been blown away by the confidence that Kraut emitted from the stage. Kraut could be delivering bold predictions for the development of nanotechnology or artificial intelligence for the next twenty, thirty or fifty years. And he would be backing up his predictions with perfectly sound arguments. Any critical questions would be shot down by a witty remark or Kraut’s uncanny ability to point out obvious flaws in his opponents’ arguments. There was a good reason Kraut had been the undisputed authority on artificial intelligence for as long as Vladimir could remember.

  And still he had been way off with his predictions.

  In Kraut’s latest book, The Morphing of Humans and Machines – The Next Logical Step in Evolution, Kraut had made the predictions that humans would be able to create a digital copy of a human brain in 2020, and that we would be able to create the first artificial intelligence on par with a human brain (AGI) in 2032. Then Artificial Super Intelligence (ASI) would quickly follow in 2033. The book had been published in early 2013, only two years ago, and at the time every single one of those predictions had been considered extremely bold and optimistic.

  Now Kraut had already missed on two of the predictions.

  On the 18th of April 2015, Neuralgo had mapped the last neural connection of Andrew Kevorkian’s brain. Two short months later the first Artificial General Intelligence had been released into the physical world by Andrew Kevorkian. The only question remaining was how far off Kevorkian would be on his prediction for the first Artificial Super Intelligence. Would it take an artificial intelligence, on par with a human brain, twelve months to become ridiculously smarter? Or would there be an almost immediate intelligence explosion? Would the artificial intelligence be capable of making smarter and smarter versions of itself a lot faster than Kraut had assumed?

  “I fucked up, OK?” Kraut said. “I didn’t see this coming. I thought I would have more time to come up with a proper response before we ended up in a situation like this.”

  “You wanted to scare them didn’t you?”

  Kraut sighed as he sat down in a chair. “Somebody had to. For a decade and a half I’ve been asking myself the question: Why doesn’t everybody talk about this? Why doesn’t everybody see what I see? And then I got this perfect opportunity. The President of the United States coming to my university. Coming to my domain. You should have seen him. It was priceless. The President was scared shitless when he got that question from my student. How could I let that opportunity pass?”

  “But you’ve been the leading authority on this subject for decades. I’ve read all your books, and I’ve attended some of your presentations. Not once have I ever heard you speak publically about the dangers of developing artificial intelligence. It has always only been about the positives – the opportunities.”

  Kraut let out a soft laugh. It almost sounded like a cough. “You know how big this industry is. Artificial intelligence powers everything from our financial industry to the stock levels in our local grocery stores. I’ve had to cultivate a public appearance over two decades so that I would be in a position to influence things before it got too bad. I thought I would have more time though.”

  “You’ve been talking up the positives all these years just to get on the advisory board for Protocol Cronus?” Vladimir asked, not believing what he was hearing.

  Kraut smiled. �
��Imagine the President’s surprise when Ronald Kraut, the biggest supporter of the development of artificial intelligence technology in the world, put the report on his desk. It wasn’t fair. I put him in a squeeze. In fact, I gave him no way out. I told him what could happen the day we invented the first Artificial General Intelligence. I told him it would most likely be our last invention.”

  “Cronus. It’s the Latin word for Cronos isn’t it? The God of Time. You wanted to stress the point that we were running out of time.”

  “No.” Kraut shook his head. “The Latin word for time is Chronus, with an h. I named the protocol after the ancient titan Cronus, the youngest of the first generation of Titans. Driven by envy Cronus castrated and killed his own father.”

  Vladimir sighed. “Let’s hope it doesn’t come to that.”

  15

  1st of June 2015

  DARPA’s remote Listening Station No 3

  The Nevada Desert

  DAY 1:

  2000 Hours

  They looked like an odd group of people, Team Cronus, where they sat in the canteen being briefed by Ronald Kraut. The soldiers standing guard by the door were clean shaven, their uniforms pressed. Vladimir sported a two-day beard and solid black rings encircled his eyes. He was wearing a green T-shirt, that read ‘Got Wood’, and his shorts looked like they could need a wash. The psychologist, Mike Hanna, had been picked up from his holiday condo in Hawaii. He had a slightly sunburnt face, and his ginger hair was covered by a New York Knicks baseball cap. He coughed constantly, having picked up a cold on the flight to his holiday a couple of days earlier. The only one who looked somewhat professional was the hacker, or computer expert, Amanda Grieves. She was a stunning petite blonde, with her hair set up in a ponytail. She had apparently been preparing to attend her brother’s wedding when the Army had arrived on her doorstep. She wasn’t happy to be there. Neither was the psychologist.