Star Trek - Lost Souls [Cortex +]
AI's in Starfleet
Due to previous incidents and encounters in Federation history (e.g. ref. stardates 2712.4, 4513.3, 4729.4) and despite the historical legal decision in 2365 regarding Commander Data’s status, The Federation Council has undertaken a cautious attitude towards the inclusion of what has been termed artificial, or machine, intelligence as a candidate species.
Fully-sentient hologram entities, officially termed HE’s but more colloquially referred to as AI’s, serve in a number of positions in Starfleet, under a few special conditions.
First, each AI is registered and tracked; this is not as onerous as it may sound, since the information is not any different than that required of any other member of Starfleet. Each AI’s holo-emitter unit is equipped with a passive transmitter that will provide the AI’s ID when queried electronically.
Another special aspect is a bit more complicated. Part of what makes an AI self-aware are complex heuristic self-modifying algorithms; essentially, they are able to learn. There is a measurable risk, though, that the algorithm may go into a cascade loop resulting in personality breakdown.
To ameliorate the risk, the algorithms include controls that limit the amount of modification they can perform. The level of control has to be carefully defined; too much, and the AI becomes just a complex holo-character program. Part of the configuration causes the controls to be lowered (as one developer described it, “easing off on the brakes”) when circumstances outside the AI’s experience and knowledge base are encountered.
As part of their development and growth, AI’s maintain a social network using a reserved subspace channel, allowing them to share ideas and thoughts. This channel is constantly monitored by Starfleet and the Daystrom Institute. Unknown to them, the AI’s implemented a piggy-back signal, a separate private channel where they can assemble without supervision to discuss sensitive topics.