Acheron Project Memos

To: Admiral Gregory L Fordeck,
SkyGuard Research Oversight Committee
From: Maj. Corbin D. Vangrace, PhD
SkyGuard Special Project Division
Subject: Project Acheron Class 1 - Initial Results and Recommendation


Executive Summary

After prototype destructive testing was complete, data analysis indicated that adding significant personality matrices to the Acheron-class AI pool was warranted. While these modifications would decrease the order of sheer intellect available to the Acheron-class AI's, they were expected to improve reliability of the model.

The first set of eight (8) post-prototype production candidates have completed initial imprinting of skills and identity generation ("Naming"). Of these, the development committee suggests that four (4) be commissioned for the stage two training program, and the remaining four (4) be destroyed.

While eight initial installations were proposed and accepted by the previous funding board, we request that construction of AI capable infrastructure halt at four of these sites, allowing a concentration of resources to develop and study the four (4) accepted candidates making up the first class.

Summary of the eight proposed candidates are below:

Proposed for Further Development:
Temporary S/NProposed Posting
ACAI-X1SkyGuard LeaveStation 1 (Hegemonic Botanical Garden)
ACAI-X2SkyGuard LeaveStation 12
ACAI-X5SkyGuard Medical Frigate AH-3
ACAI-X7SkyGuard Starship Construction Facility 15
Proposed for Destruction:
Temporary S/NProposed Posting
ACAI-X3SkyGuard DEW Outpost 1
ACAI-X4SkyGuard Weapons/Ordinance Research Division Station 1
ACAI-X6SkyGuard Tactical Training Academy Godia
ACAI-X8SkyGuard Academy Ganfrey

To: Maj. Corbin D. Vangrace,
PhD,SkyGuard Special Project Division
From: Admiral Gregory L Fordeck
SkyGuard Research Oversight Committee
Subject: Project Acheron Class 1 Oversight Decision


Executive Summary

The Research Oversight Committee has elected to continue the Acheron project with all requested funding.

The committee makes the following response to the recommendations of the research team.

ACAI Temporary S/N
ACAI-X1:Recommendation Accepted: Proceed with phase two
ACAI-X2:Recommendation Accepted: Proceed with phase two
ACAI-X3:Recommendation Rejected: Proceed with phase two
ACAI-X4:Recommendation Accepted: Proceed with destruction
ACAI-X5:Recommendation Accepted: Proceed with phase two
ACAI-X6:Recommendation Accepted: Proceed with destruction
ACAI-X7:Recommendation Accepted: Proceed with phase two
ACAI-X8:Recommendation Accepted: Proceed with destruction

To: Maj. Corbin D. Vangrace, PhD,
SkyGuard Special Project Division
From: Admiral Gregory L Fordeck
SkyGuard Research Oversight Committee
Subject: Acheron Class Recommendations - Clarifications Requested


Corbin,

Can you get me some details on the committee's recommendations?

Particularly with respect to ACAI-X3 and -X4? There's a lot of resistance on the oversight committee to abandoning those two as their programmed specialties were of great interest to my superiors. And I don't need to tell you how many superiors I have.

Get me some hard facts, or I won't be able to swing your recommendations with the committee.

-G.

To: Admiral Gregory L Fordeck,
SkyGuard Research Oversight Committee
From: Maj. Corbin D. Vangrace, PhD
SkyGuard Special Project Division
Subject: Acheron Class Recommendations - Clarification of Destruction Candidates


Admiral,

Here are summaries of the research teams. Complete profiles will follow under a separate cover.

ACAI-X3: This candidate chose the name "Deep Blue". The research team feels uncomfortable with this candidate as initial extrapolations already indicate some tendency towards emotional response matrices correlated with arrogance and resentment. The fact that it named itself after a machine rather than using a human name leads to some indication of the core instability of the Acheron type.

ACAI-X4: This candidate chose the name "SkyNet". Though the research team was initially encouraged on discovering that the name referred to a fictional machine rather than a historical one, indicating increased imagination, once the actual details of the genre fiction were revealed, the team felt immediate deactivation was in order. Considering its proposed posting, I think the reasons for this are clear enough.

ACAI-X6: This candidate chose the name "Icarus". The research team feels the name indicates a long term outlook inappropriate for a SkyGuard Training Academy. Perhaps the initial skill set was not properly calibrated for its expected duty load. There is significant dissent among the team on this one though, and the vote was close, so proceeding with this candidate would certainly be possible. It was included among the deactivations because the vote was close and we are trying to be sensitive to resource allocation.

ACAI-X8: This candidate chose the name "$name". At first, our team thought that name selection had simply failed. However, it seems this candidate actively chose this variable as a name. This is wholly inconsistent with the programmed acceptance of the personality matrices. Extrapolations of future emotional responses are mostly flat. We feel the candidate doesn't have the core Acheron flaw, but our attempts to instill mediating personality matrices have resulted in an inconsistent design, thus the name selection anomaly. We expect performance would be significantly sub-par if activated.

To: Maj. Corbin D. Vangrace, PhD
From: Lt. Darlene Case, PhD
Subject: Dissenting Opinion


Major Vangrace,

I feel I must renew my dissent with the recommendation report forwarded to the committee.

I understand that my views are in the extreme minority among the research team, but I must again point out that if we are incorrect, we could be making a potentially life threatening mistake.

Gender is simply not a Hermetian operator across the personality matrices. The fact that candidate "Xanadu" chose what "she" claims is a female name and personality basis when built upon a clearly male template is a clear sign of possible instability. And don't try to pawn me off with, "but it was a gender-neutral template" as Dr. Lal is constantly saying. As the only female on the team trained in that type of advanced psychology, I'd think you'd take me seriously when I insist that the initial load-out of all eight databases were disticntly male.

We need to meet and discuss this issue ASAP before the committee makes its recommendations.

-Dr. Case

To: Dr. Perseid Lal, PhD
From: Maj. Corbin D. Vangrace, PhD
Subject: Unauthorized Research


Dr. Lal,

I am instructing you to terminate your simulations with the Acheron AI class regarding prototype reconstruction from viable production candidates.

Even though the Acheron prototype possessed a higher order intellect than the current production model, all tests showed that such an intellect was basically unstable. The ACAI-XP was decommissioned precisely because of this instability, which led to the events of which you are well aware, and simulations on how to re-create it are counter-productive.

I am logging this order in your personnel file and expect all simulations materials to be immediately terminated and transferred to deep storage.

-Major Vangrace

To: Maj. Corbin D. Vangrace, PhD
From: Dr. Perseid Lal, PhD
Subject: SCIENCE!


Major,

You, the committee, and even the Strategos himself are all failing to see one crucial point! There is a great difference between "Immoral" and "Amoral". Just as there is a difference between "instability" and "unpredictability".

A higher order artificial intellect is unpredictable precisely because it is of a higher order! You simply can't discontinue the original design goals of the project, in order to satisfy the expediency of the military oversight!

First you wouldn't let us keep working with the original prototype! Now you are classifying my technique to recreate the intellectual order of the original prototype by combining several members of the functioning production class! Where's the SCIENCE!

Dr. Lal

To: Lt. Darlene Case, PhD
From: Maj. Corbin D. Vangrace, PhD
Subject: Secrecy


Lieutenant,

I have considered your recommendation carefully and have decided to proceed as originally discussed immediately after the termination of Dr. Lal at our weekly staff meeting.

Turning your argument around, you are the one who is constantly pressing us to treat them with the consideration and respect we would give humans. Well, I can tell you right now, were they humans in this situation, I certainly would not tell them that they were the only five survivors of an eight member class, considering that they didn't know each other ahead of time. It would be different if they had known each other, but in this case, my original decision will stand.

Aside from basic morale issues, it would undermine their confidence and thus their ability to pass the very tests you wish to prepare them for in order to prove their sentience.

So, as previously instructed, run the computer search to find and classify all documents referring to the original numbering scheme, transfer them all to deep storage, renumber the Class 1-5, and proceed with phase two commissioning.

-Maj. Vangrace