China: Technology
The Tiāngōng Intelligence
"This is one of the pitfalls of fighting a less advanced, but intelligent species. It does not take long before they take our strengths, and make them their own."
- Ethereal Fectorian
MINISTRY OF STATE SECURITY CONFIDENTIAL
To Agency Chief Chen Yongyue
Agent: [REDACTED]
Our engineers have finished our own assessment of the remains of the fallen alien Dreadnought, as per the request of the Ministry and President Qin's own instructions. While it was heavily delayed thanks to the deployment of United Nations Peacekeepers to hold the remains while the bureaucracy was sorted out, they have left and there was much ground to make up for the lost days and weeks.
While there was clearly some attempted sabotage (Likely by XCOM) on some of the more valuable components we tried to recover, the vast majority of the ship is salvageable, including the alloys from which the ship was created. We should work to integrate those either into our armor or aircraft programs as this could give us an incalculable advantage as the current political situation intensifies.
The vast majority of computers were also destroyed, and the ones that we managed to restore some power to are unresponsive and in some incomprehensible language. There is no direct means of interfacing with them that we could identify, though this is still ongoing. The alien corpses recovered are surprisingly few, but they will likely provide some use to our science teams.
But there is a surprisingly good piece of news.
XCOM overlooked something major; something I am personally shocked they did not encounter either on the ship, or in the immediate aftermath. I can only presume that this oversight was due to them being rushed and getting as many obvious useful components and sabotaging the rest before being forced to leave.
To explain, there was a kind of central computer deep within the ship. We did not know it was a computer at first, since it was a supposedly sealed black box with notable connections to nearby power systems. I presume XCOM took that to be another power system, or not even something worth looking into at all.
But this system, it turns out, had some kind of artificial intelligence within. Upon our engineers breaching the central computer box, it activated of its own volition. It appeared as a golden glowing humanoid figure; faceless and genderless initially. It took a few moments for us to realize that it was waiting for something, so the engineers began interacting with it.
It appears to have all major Human languages stored within it, and the working theory at the moment is that the crash somehow managed to 'reset' it to a factory state. There is sadly no data that could be salvaged by the intelligence, since the computers it was connected to were destroyed. We are still as in the dark about the aliens as the rest of the Council.
However, it seems willing to assist us wherever we need it. I'm sure you can see the potential of this. If we could transport it to a more secure location, or replicate it, not even the United States would likely be able to interfere in our work. However, care must be taken as if we do decide to withhold this from XCOM, I highly doubt they would take it well.
I might suggest considering sharing some details. Not the full scope, but enough for them to trust us. Despite their allies delaying our rightful claim over the wreckage, I see no point in keeping them entirely in the dark. They do not seem a wise enemy to make.
However, the decision is up to you.
ADVENT Report 2211
Subject: Chinese Tiāngōng Intelligence
Author: Commander Laura Christiaens
Recipients: Chancellor Saudia Vyandar, Chief of Lancer Operations Helion Weekes, Chief of ADVENT Engineering and Development Feng Mercado, Chief Diplomat Firdaus Hassan, Director of ADVENT Intelligence Elizabeth Falka, XCOM Central Officer Ariel Jackson
We had, of course, suspected that the Chinese were intentionally holding something back given their insistence on taking charge of the operation to liberate Japan. I suspect the initial goal for their secrecy was to create a sense of awe and surprise in us, perhaps making us question just what else they had waiting for them.
That has appeared to not take into account that we are far more advanced than them in almost every way, but we will ignore that for now, suffice to say this is not likely how the Chinese predicted it turning out.
With that said, if ADVENT did not exist, this would be the most consequential development in the realm of artificial intelligence in decades, perhaps of all time. The Tiāngōng Intelligence, as the Chinese have designated it, is at the moment the most advanced machine intelligence in the world. The Chinese have fortunately been courteous enough to share some details as to how the intelligence operates.
The first thing that you should be aware of is, contrary to some claims, the Tiāngōng Intelligence is not a true Artificial Intelligence. It is a highly advanced machine intelligence, but it does not have a set personality or motivations beyond its programming. True artificial intelligence development will no doubt be improved by the Tiāngōng Intelligence, but it is important that we do not confuse the different types of computer intelligences.
The second is how the Tiāngōng Intelligence actually works. The purpose is to serve as a battlefield micromanager for each individual soldier, and give specific and effective orders in conjunction with other soldiers at the same time, which turns even a smaller and weaker military into a vastly superior force.
This is a level of coordination that is flat out impossible for any organic commander, but it is something that the Intelligence seems surprisingly good at. The Chinese have devoted, in addition to the internal CODEX infrastructure, every supercomputer they have developed (Over 200) to assist in processing speed to handle the sheer amount of information.
They have informed us that they are working to improve the processing speed, so that it doesn't require every piece of computing power they have when the intelligence utilizes its full potential (Such as with Japan). In fact, the liberation of Japan served as a massive stress test on the Intelligence – one which it passed, although during that time anything even remotely connected to the Intelligence slowed to a halt.
There are still clear kinks to be worked out in the system, but the fact remains that this is a major step forward for this kind of battlefield intelligence, it is without a doubt something that we need to look into emulating. Having our entire military under one of these battlefield intelligences is not advised, but one for individual legions is perhaps something to look into.
I would also caution that we do not become reliant on something like this. While the Japan liberation was a success, it is, in the end, a machine that can only adapt as far as its programming allows. An ADVENT battlefield intelligence is something that should be another tool in our arsenal, not a crutch. Much like psionics, if you wish to make that comparison.
Another thing we need to take into account is that if China is able to utilize something like this, the aliens can certainly do so as well – and can likely do it better. I would not be surprised if they attempted to emulate this, in the event that something similar does not exist. We should prepare for how to defeat a similar system, even with what limited information we have.
Joint mock war exercises with the Chinese under the Tiāngōng Intelligence might be a good way for us to prepare, and the Chinese to refine. This is, of course, in conjunction to what China has given us to make our own Tiāngōng programs. We will have more on that as the project develops further.
CODEX File ****
Subject: The Battlefield CODEX
Location: Blacksite ***
Author: Engineer Fectorian
It has become somewhat irritating seeing another species use fairly obvious ideas and applications of our own technology before we have done so ourselves. The fact that this was not done by a major Human power like ADVENT or XCOM, but China is somehow more insulting since by all sentiments they were not really supposed to have that capability.
I have personally suspected that ADVENT would be exploring the possibility of artificial intelligence, and this recent incident has only served to confirm that it will come eventually. The Tiāngōng Intelligence, as the Chinese call it, is a solid amateur step forward. I suspect that with the backing of ADVENT, they will make an intelligence that is somewhat more complicated.
Taking data and giving pre-determined commands works as a simple enough concept, but it is all algorithms created by fallible beings. I suspect that there were many instances where this intelligence got scores of soldiers killed, but in the scope of the entire battle, they were either not noticed or not worth mentioning. The Tiāngōng program is not well-optimized, that is clear both from its actions and observing energy readings from China itself.
It will improve, but the flaws will become very apparent if this is used more frequently. The Chinese appear to have been blinded by their shiny pieces of technology, at least from their laughable action to put their entire military under the direction of it. Questionable, and a major risk if they keep developing it. Perhaps expected with a species with no experience in AI research, but still remarkably short-sighted.
Nonetheless, it is something that can be answered with ease.
The CODEX system is already primed for such usage. The Chinese have adapted it crudely. I believe it is time for this potential to be fully unleashed. The Battlefield CODEX concept has sat and languished among the sadly long list of improvements to be made, and now with the Battlemaster finally taking steps to address the deficiencies in the Collective, the time has come to unveil it.
I suspect Cogitian will not be pleased at the further weaponization of his system, but he got over his archival system being used more broadly for data analysis; he will get over using it to improve our military. It would likely be a decent idea to bring on further experts on this, given the new emphasis of inter-species relations. Union Apear supposedly specializes in this, and I am sure they would deeply appreciate the opportunity to observe and improve a CODEX.
One more project to begin, but this is another that has the potential to change how war is waged. And that I look forward to. Whatever can be said about the Humans, they have forced us out of stagnation, and for that they should be thanked.
Internal Report Sent to the Commander
Commander,
I have been reviewing the material regarding the Tiāngōng Intelligence that the Chinese provided us, and while we do not have a replica up and running right now, from the interviews, reports, and looking over the material myself, I can already see how we could begin applying this to several of our own projects.
I do not necessarily believe that we need such an intelligence for our own soldiers. This works well in larger armies, but would become unnecessarily redundant in the squads we so typically deploy. There is also the fact that we would need to rework our existing armor and introduce potential vulnerabilities to ensure it is safe to use. There are also arguments to be made on the wisdom of relying on a machine intelligence, no matter how advanced.
However, there are two major areas where we can apply the information that we have now:
Defense: We know everything about the Tiāngōng Intelligence, inside and out. With this we will be able to plot logical deficiencies, predictive algorithms, and how the intelligence gives specific orders. We already know how power-intensive the Intelligence is, as well as how computationally draining, and while those are obvious vulnerabilities to exploit, they are not the only ones.
In the end, this is not an actual AI. It is driven by Human-made algorithms and code. Humans make mistakes, and even if they did not, all we need is to figure out how the Intelligence "thinks" and then manipulate it to achieve a desired result. From looking over the major parts, it does not seem to be able to "learn" as you or I define it. Everything has to be tweaked by software engineers or verbally. It does recognize if a mission has failed or not, but it does speak to how much the Chinese are relying on this where it is barely more complex than that.
They do not expect this to fail, or at least they expect casualties and results to fall into acceptable ranges.
The point to make here is that the Collective will almost certainly try and emulate this at some point, and we can only assume it will be better. We need to come up with ways of neutralizing or negating an advantage like this, and making smaller versions of the Tiāngōng Intelligence to test on would be a valuable usage of it.
JULIAN: Obviously, we are working on our own AI project, and this provides us with a significant step forward. Much of the "personality" parts of the Tiāngōng Intelligence are…muted, for lack of a better word, or disabled entirely. The Chinese did not seem to want to create a true AI, but they have ironically made it much easier for others to do so since they have thoroughly gone through enough of the code to note that if they did want to embark on such a project, you would modify these specific parts.
A true AI must be able to modify its own code, to a certain degree. This has been something of a problem to figure out at least in determining boundaries. We have been able to see prior CODEX units can do this to an extent, it is how they "develop" according to Aegis to be more effective, but we were still working on reverse-engineering the code itself. Fortunately, the Chinese have done all that for us.
With a more comprehensive understanding of how CODEX systems work, we can continue to adapt the JULIAN program into an actual AI at a far faster rate than we have been so far. A true AI will be able to provide all the advantages of the Tiāngōng Intelligence, and will more than likely be much better at doing so.
While Aegis does not approve, I believe it's something that is still worth pursuing. We just need to ensure that all proper precautions are taken. Thanks to several of the Andromedons, such safeguards will be the best that can be conceptualized. I would also suggest we send our thanks to the Chinese as well, regardless of how…forcefully we took it. But it never hurts to smooth things over when appropriate.
- Raymond Shen, Head of XCOM Engineering