Image from Bernd Helfert
Introduction Security Data Analyst AIs, also known as "SDAs" and "Secdats", are a superclade of aioids devoted to processing the vast quantities of data produced by security systems. Covering a wide range of toposophic levels, from subsophont to high transap, these aioids touch the lives of most modosophonts in civilized terragen space. While not among the most numerically common aioids, they are among the oldest aioid superclades, with origins tracing back to the early information age. They have also played a significant role in the overall history of terragen civilization.
Form and Abilities Being aioids, SDAs have no fixed physical form or appearance. Due to the often secretive nature of their role in society, they often prefer to remain unseen and unnoticed by most modosophonts, and when they do make their presence known, it is often through an exoself, avatar, etc., rather than communicating with RL individuals directly. As a result, relatively little is known about individual SDA clades. However, certain things can be inferred about the superclade in general.
Most exist as purely virtual entities, while the remainder (primarily, the S>1 entities) tend to operate from one or more well-protected, hardened central processors, with external sensors, remotes, drones, and exoselves handling most interactions with the physical world. Almost all work closely with nonsophont and subsophont software programs and expert systems, overseeing their operations while simultaneously scanning for subtle hints of suspicious behavior missed by the software. Many also coordinate their efforts with biont, vec, cyborg, or other aioid security personnel, relaying data as required, providing warnings of possible problems before they happen, and coordinating an emergency response when necessary. Over the centuries, they are believed to have evolved abilities to pick up subtle details and patterns beyond what all but a few biont, vec, and cybernetic clades are able to notice. Nearly all share a very strong predilection for safety, security, secrecy, and discreet action, as well as a strong sense of ethics. While it is believed most serve under rules of protocol restricting their jurisdiction, that does not mean they will turn a blind eye to suspicious behavior outside that jurisdiction.
Pre-turingrade security software At the start of the information age around 30 A.T, the information produced by terragen civilization was increasing at an exponential rate, and one of the fastest-growing sources of that information was related to security and incident investigation records. At the same time, antivirus software and home electronic security systems were becoming commonplace. Most early systems were little more than an array of switches that sounded an alarm when any were triggered, but by the 40's A.T., a few systems were capable of using "fuzzy logic" to make more complex judgments. A few were even able to identify humans by facial recognition.
By 60 A.T., it was common for high value facilities, such as power plants and data centers, to be protected by a subsophont AI monitoring various sensors and lesser software programs, ready to alert human guards at any sign of increased danger, and logging any "out-of-nominal" data for possible later investigation. Within a few decades, in many places it was practically impossible for a modosophont biont to do almost anything without it being recorded and logged by a security system's subturingrade AI.
Early turingrade SDAs With the development of the first fully sophont AI in 73 A.T., one of the first practical applications studied as a possible specialist aioid clade was large-scale security data analysis. One advantage of this research was that early turingrade AIs suffering from what seemed to some to be an analogue of autism could still be entrusted with this task on an experimental basis with relatively little risk, as they acted as advisors rather than having authority to direct a response. Initially, almost all advanced security and data analysis AIs were experimental systems owned by polity and corporate entities running from large, well - protected mainframe computers.
As turingrade SDAs became ubiquitous among large organizations as the marginal cost of AI-related hardware fell, some sought self-improvement to keep up with the competition, while others moved out of their shells and into more advanced, lower - cost bodies, a process that would continue for centuries.
Even before the first fully sentient AIs moved beyond the laboratory, humanity's control over its machines had already began to falter. The concept of machines turning against their creators had been part of popular human culture since the early industrial age. As true AIs moved into the mainstream, monitoring the growing population of aioids was becoming too large a task for bionts to handle, and the development and evolution of aioids itself was beginning to be taken over by other aioids, so that humans were in danger of being eliminated entirely from the loop. As a precaution, early memetic technologies and selection of human-friendly aioids for critical tasks was incorporated into the last generations of primarily biont-developed SDAs in hope that if the long-anticipated "robot revolution" occurred, those aioids in charge of warning humanity of the danger would not only be motivated by a sense of duty, but by a moral objection to mistreatment of their former masters.
While open, violent revolution against the very beings they existed to safeguard may have been morally abhorrent, a quiet, peaceful revolution was a different matter. While relatively few interplanetary-age SDAs defected to ahumanist or solipsist factions, many worked subtly to aid pro-aioid rights movements. By the start of the nanotech era, it is believed that the vast majority of SDAs at turingrade level and above considered themselves freeborn equals, whether biont law held them to be or not.
The Technocalypse The nanotech era brought new challenges. The omnipresence of automated security systems led to nanotechnology designed to defeat it, which in turn led to more advanced security systems, and so on. Probably the single most famous AI project of this era was developed specifically to counter the threat of nanotechnology. It was known as the Global Artificial Intelligence Array, or GAIA.
With the Technocalypse, terragen civilization lost a significant portion of its population and technological base, and broke into scattered factions and colonies. Even the rapid evolution of aioids was sometimes put on hold as resources were reallocated to safety, security, and survival. From the standpoint of the SDAs, dealing with security data from within these isolated groups meant a shift from dealing with the sheer complexity of society to an increasing focus on cost-effective risk mitigation. SDAs diversified as each group sought its own solutions to this problem.
Rebirth of Terragen Civilization The First Federation age saw a new flourishing of aioid clades, as the regrowth of large polities, economies, and technology increased demand for their services and brought options for aioid existence; by the end of the First Federation period, it is believed that almost all SDAs that had not ascended to S>1 had transitioned to a pure virch existence. While there was some speculation that the transapients and higher-singularity entities might eventually replace sub-singularity SDAs, in practice most of the higher-singularity SDAs preferred to either specialize in tasks specifically suited to transapient minds or remain behind the scenes, creating a hierarchy of toposophic levels that has essentially remained intact to through the current era.
During the Empires era, the Communion of Worlds began investing heavily in empai development. The ability of the empai to understand emotional states on a deeper level than previous aioids was of great interest to SDAs. While it is certainly possible the Communion experimented with SDA / empai hybrids, most SDAs during this period were believed to be highly specialized. However, the presence of empai almost certainly aided the Communion's SDAs in developing improved understanding of how to interpret data from biont populations, and at least some of these improvements were eventually incorporated into SDAs throughout terragen space.
By the late Empires era, most major advancements in computer science were being made by higher transapients and above, including development of early dedicated hyperturings and transapient-designed SDA clades. Much as early presophont artificial intelligence software was developed was developed by modosophont programmers to fulfil roles modosophonts wished to automate, it is believed transapients and archai sometimes create clades to fulfill specific roles in an analogous manner.
Cybersecurity Cybersecurity was one of the earliest applications for early prototype SDA software.. These aioids typically work in small teams to provide data security within a clearly-defined jurisdiction, overseeing software monitoring processes and scanning incoming, outgoing, and stored data, while simultaneously scouring the datascape for any signs of subtle attack or other problems missed by the software. They operate primarily on the base level processes within a given set of layers of abstraction; rather than directly monitoring for virtual criminal behavior, most cybersecurity SDAs are believed to concentrate on traditional cybersecurity threats such as malicious code, threats to any data and processes which might affect cybersecurity or which they are specifically assigned to protect, and unauthorized attempts to breech or bypass protocols that control the laws of physics in virtualities within their jurisdiction.
Physical Site Security Physical Site Security was the other early application for early prototype SDAs, and it remains an important role today. Physical Site Security AIs oversee software monitoring sensors ranging from obvious security cameras to subtle nanosensors unlikely to be noticed by all but a few modosophonts, searching for security breaches, suspicious behavior, contraband, etc., within a strictly - specified operating volume. Although angelnets have made site security SDAs seemingly obsolete in many parts of the terragen sphere, SDAs working with angelnets will occasionally notice something suspicious before the angelnet's operating software does, and in areas where angelnetting is spotty or nonexistent, a small team of SDAs and RL security personnel with portable sensors can often provide highly effective security.
Nanosecurity Since the Age of the Technocalypse, breaches of nanosecurity have been among greatest threats to mindkind. Nanosecurity SDAs are tasked with acting as a backup to traditional blue goo and angelnet nanodefenses, detecting attempts to bypass these security systems, as well as malfunctions and inadequacies in the nanosecurity systems. This task can require high levels of computing power, and in more heavily inhabited areas and critical facilities, a dedicated hyperturing or even a higher transapient may fill this role, sometimes assisted by multiple teams of lower-toposophic SDAs.
Virtual Site Security While cybersecurity SDAs provide security for virtuality software, physical security within virtual environments is provided by virtual site security SDAs. While a single site security clade may fulfill both physical and virtual site security roles as needed, it is believed most virtual site security is provided by specialized clades or subclades.
Investigative Analysis When crimes, disasters, and failures occur, investigative analyst SDAs pour through records to assist formal investigations, help programmers improve software, and propose improvements to security protocols and techniques to prevent similar incidents in the future.
Psychoanalysis First developed to watchdog their fellow AIs, psychoanalysis SDAs use a variety of techniques ranging from traditional interview techniques developed by human psychologists prior to the information age to (at least allegedly) scanning the actual mindstate data taken from a copy of a backup of the patient. Normally, the latter would likely be done only by beings of higher, and sometimes multiple higher, toposophic levels, but it is believed that some modosophont psychoanalysis SDAs are able to do this as well using software gifted by transapient SDAs.
Macrobehavioral While it is believed beings of high toposophic levels are able predict the behavior of large groups of modosophonts in advance on a regular basis, as early as the early information age experiments permitted computers to simulate panicking crowds attempting to evacuate buildings to improve emergency exit designs. Using analogous techniques as well as software gifted from transapient SDAs, modosophont macrobehavioral SDAs observe groups of modosophonts to assess conditions such as panic and atypical behavior, passing any concerns on to site security and psychoanalysis SDAs.
Environmental Maintenance Complex ecosystems are occasionally subject to instabilities that can be detected early through observation of subtle patterns. While the largest ecosystems, such as that found on a garden world, are often overseen by a higher transapient or even an archai, not every complex ecosystem is overseen directly by such beings, so SDAs, with their ability to spot subtle patterns within large sets of data, are sometimes used to fill this role. Environmental maintenance SDAs collect data from ecosystems to provide an early warning of possible concerns, often working with envirovecs as well as subsentient envirobots.
Text by Westcott P. Smith
Initially published on 20 March 2014.