The Defense Advanced Research Projects Agency

675 North Randolph Street

Arlington, Virginia 22203

Eugene Hubertus Turing

Professor of Artificial Intelligence and Robotics

Curry College

Hanover, New Hampshire 03755

Dear Professor Turing,

We at DARPA believe the C.A.R.O.L.I.N. Project holds great promise for the future security of our country. We look forward to merging your achievements in Artificial Intelligence with our knowledge of robotics. Machines that can work together while displaying human tendencies are the next logical step in the defense of our great nation, and the world-wide peace that defense provides for all humanity. Using a computer network such as C.A.R.O.L.I.N. to operate a team of robots means that the men and women who serve in our armed forces, protecting us from enemies both foreign and domestic, will no longer have to be placed in harm's way in order to do their job.

Developing robots that can work together is of great importance. They must be able to respond in a real-time manner to unforeseen changes in their environment. Housing the computer that runs these robots at a remote location, far from harm, means the program that runs the C.A.R.O.L.I.N. Project will be kept safe and secure, stabilizing the entirety of the system. It will allow these teams of robots to successfully complete their mission, regardless of damage or loss.

With the computer network operating this system located far from combat, it's possible to build an android shell that is resistant to damage, and thus less prone to failure. These android shells will be integrated into the computer network using an array of servers and other wireless devices, giving them a range of freedom and movement that is unsurpassed, and does not endanger the system itself.

We've already constructed a prototype of one such android shell. It, and many more like it, can be operated by a computer system such as your C.A.R.O.L.I.N. Project. The shell has built-in redundancy, affording it the promise that it will still function at optimal capacity despite a level of damage or loss that can near ninety percent. Furthermore, we hope, should such damage arise, C.A.R.O.L.I.N. will accurately assess the loss and instruct other machines in the field on how to affect repairs.

A computer system that can assess its own damage and affect repairs upon itself, all the while responding to the demands and needs of a hazardous environment, give us a significant advantage over an enemy in the field.

Professor Turing stopped reading for a moment. "You meant to say 'weapon system' here," he said to the man from DARPA. "Right?"

"Excuse me, Professor Turing?"

Turing looked him square in the eye. "You say here that my Project will be operating a computer system. You meant to say that it will be operating a weapon system. Right?"

"Professor Turing…" Dean Dayne began, preparing to chastise him.

He didn't give her the chance. "Why can't we get somebody interested in the Project who doesn't want to use it to kill people and blow stuff up? Why is it always the military?"

"The military has deep pockets," Dean Dayne replied with firmness. "Your project costs a great deal of money. It has high overhead."

"Because it's capable of so much! Moreso than just killing."

Haskins spoke up. "Professor Turing. Sometimes the role of the military is to prevent further bloodshed. A military system…"

"A weapon system." Turing interjected while Haskins continued talking.

"…controlled by your C.A.R.O.L.I.N. would actually save lives. It would save the lives of those who are doing dangerous jobs right now. Those who are in harm's way. And with a battle-hardened android shell, it will survive against a high level of enemy resistance, whereas a human being might not."

Professor Turing stopped listening to Haskins after about a sentence in order to finish reading his cover letter.

We feel certain that the C.A.R.O.L.I.N. Project can operate at peak efficiency in these conditions. To guarantee that with certainty, the Project must be tested under 'partial system loss' scenarios. Successful test results using our android shell will provide us with the data we need to confirm that your Project is capable and effective while under duress. Once the C.A.R.O.L.I.N. Project has demonstrated what a team of networked robots can do, it will become a permanent link between machine reliability and the human decision-making process.

We look forward to working with you, and hearing from you soon, in regards to when these tests can be implemented. We have every confidence that your C.A.R.O.L.I.N. Project will achieve excellent results.

Godspeed, Professor Turing.

Most respectfully yours,

Robert Haskins

AI Project Manager, Robotics Division, DARPA

Godspeed indeed, Turing thought, sighing to the inevitable.

"Eugene," Dean Dayne calmly said, "without a budget, there won't be a C.A.R.O.L.I.N. Project at this college. Other institutions will take what you have done for us, and once they have achieved success, they'll take the credit for the many years of hard work you've put in."

The author's tale has been misappropriated; report any instances of this story on Amazon.

Professor Turing said nothing. The table grew silent as he glared into every eye.

"Professor Turing," Chief Budget Officer Aaron Hilbert said after a time. "We have other research projects with DARPA at Curry College."

"I'm well aware of that."

"Well, then, you must know about the research we're doing in the field of Functional Analysis."

Turing blinked as he struggled to recall some knowledge about the field Hilbert had mentioned.

"Behavior Recognition software," Hilbert furthered, jogging Turing's memory.

Turing brightened up a bit upon hearing those words. Behavior Recognition software was the next step in voice recognition software. Not only could words be understood and analyzed by a computer using this software, but also mood, inflection and inference. He knew the research in this field was moving at a fast pace, and going on competively at other campuses around the world. Unlike the cost of the C.A.R.O.L.I.N. Project, Functional Analysis had a low overhead. Also unlike C.A.R.O.L.I.N., a breakthrough in the field was due at any moment.

Turing spoke as if he were giving a lecture on the subject. "Functional Analysis applies flexible conditional laws to the decision-making process a computer program uses. Motivation, behavior, attitude and consequences are analyzed before a response."

"That's correct, Professor Turing," CBO Hilbert said. "And I believe the results we've achieved in regards to Functional Analysis at Curry College would be of interest to you."

During the pause that followed, Hilbert turned to the man from DARPA. "Mr. Haskins, Functional Analysis is one of your projects. Perhaps you'd like to tell our colleague what we've accomplished recently."

All eyes turned to Haskins, with no one showing more curiosity than Turing.

"We have a reliable and well-tested system, where voice recognition software is melding seamlessly with behavior recognition software. Together, they are providing an accurate interpretation of any spoken word or written sentence. Is the speaker being sarcastic? Are they confused, demanding, cynical, frightened? Functional Analysis derives clues from things such as pitch and tone, inuendo, inflection and word choice, but then goes one step further."

"How so?" Turing asked, intrigued.

"It examines the physical orientation of the speaker. It analyzes body language, gestures, expression. Is the speaker making eye contact? Does he seem hesitant, or unsure? Is his delivery hurried? FA software provides an outline to the circumstances motivating the speaker to say what he or she is saying.

"Finally, our FA software at Curry College has taken a crucial step, and can now assess the likely outcome of various different responses to a perceived behavior. Which response best serves the speaker? Which best serves the situation? Will it help the speaker if he is in need, or thwart him if he's dishonest? What are the ramifications in regards to right and wrong, fair and unfair? Will the response be of benefit to others, or will it affect them adversely?"

Haskins paused for dramatic effect, sensing Turing's fascination.

"Our Functional Analysis Behavior Recognition software is available now, It can do these things, and more. The only thing holding us back is embarking on a phase of implementation." Haskins inhaled to further pique Turing's interest. "We need a computer system worthy of having our software installed."

Turing spoke at a near whisper. "The C.A.R.O.L.I.N. Project, you think?"

Haskins nodded sincerely, a smile on his face. "With this software installed on your project, perhaps these lockups you are experiencing will all go away. You'll be able to talk to C.A.R.O.L.I.N., as it has synthetic voice capability. Perhaps then she'll respond better to your commands. She'll learn to behave, and give better results."

"How would I get this software installed? Who would implement it?"

Some paper shuffling and throat clearing gave Professor Turing the impression that, for some reason, this question wasn't easy to answer. Dean Dayne then spoke up.

"Eugene, you and I, we've been over the subject of how to fund the C.A.R.O.L.I.N. Project many times. I've driven myself to near insanity, floating funds that ought not be yours, hoping—and sometimes praying—that money will come from somewhere."

Eugene pursed his lips and nodded, agreeing with what Mandy was saying. To that, she gave a warm smile.

"Sometimes it's been necessary to borrow money from the Functional Analysis people. Though you may not know this, Mr. Taylor does."

"How?" Eugene interrupted, feeling a bit perturbed.

"Because, unlike you, he talks to people involved in other projects on campus. He has friends."

"The C.A.R.O.L.I.N. Project takes up a lot of my time," Eugene said weakly, in his defense.

"Indeed it does. And at times, when you're busy working out some kink or whatever in C.A.R.O.L.I.N., Mr. Taylor looks around at other projects we have going on—like Functional Analysis and its Behavior Recognition software. It's a promising field. It's achieved laudable results."

Eugene head rose as he straightened his posture in response to what Mandy implied. "Geoffrey wants to work on Functional Analysis," she said, "rather than on C.A.R.O.L.I.N. And if Mr. Taylor goes, so will Mr. Lucas, leaving you with no grad students working on your project. No grad students means no funding for your Project, Eugene. But if we were to merge the two…"

After a heartfelt sigh, Eugene asked the next logical question. "So, how much money are we talking about, ah…" He searched his cover letter for the name of the man from DARPA.

"Haskins, Professor Turing," the man replied. "Robert Haskins. As you can see in our report, based on a three year history of your expenses, we can match that and more, with an increase in pay to your staff. We'll also supply the funds you need to install and test our Funtional Analysis Behavior Recognition software, and our android shell."

"Your weapon system," Turing muttered to himself, loud enough to cause Dean Dayne to inhale audibly.

"A robotic system with a battle-hardened shell. Yes sir. That's right. It's not our intention to deploy these devices as some sort of weapon system. They're far too costly, and it's too dangerous to risk letting them fall into enemy hands. Most likely, C.A.R.O.L.I.N. will act as a scout. It will perform reconnaissance and find weaknesses in our enemy, and ways to break through their lines. It will create a clear path for our brave men and women to move freely—and safely!—in the field."

The table grew silent while Professor Turing weighed his response. He then spoke to Dean Dayne.

"Will Taylor and Lucas help me install this battle droid? I seriously doubt they will. They're no more interested in weaponizing C.A.R.O.L.I.N. than I am."

Dean Dayne spoke sharply, yet as a friend. "That's why we're not installing a battle android, Professor Turing. It's just a shell, and it's only the torso. With it, your C.A.R.O.L.I.N. Project can respond to commands in a more human-like fashion." She leaned in for effect. "That is the goal you're trying to achieve. Is it not, Eugene?"

With silent resignation, Turing agreed to the conditions that had been laid out before him. "When will we install this shell?" he asked. "And the recognition software?"

Robert Haskins body language spoke of his excitement. "Right away, Sir! Everything you need on our end is at the ready. All we need is C.A.R.O.L.I.N."

My precious, dear-to-me Project.

Dean Dayne began wrapping up the meeting. "So we have a deal? Everything is agreed?"

"We have a deal," Eugene said with little joy.

Please forgive me, C.A.R.O.L.I.N.