Welcome to Real Good AI's Digital Commons feat. Mark, Bob, and Wade
A Speculative Horror Writing & Audio Podcast Competition by Real Good AI
Welcome to Real Good AI's Digital Commons feat. Bob, Mark, & Wade
A liminal space of smoke and algorithms, between the real and the virtual, tales of machine dreams, imagined by human minds.
A Speculative Horror Writing & Audio Podcast Competition by Real Good AI
FAQs
-
Yes, yes it is.
-
During the July 6 - July 17 submission window, the website will prominently feature a link to a Google Form you will need to fill out. This includes a few administrative questions, will have a place to upload your primary document as a pdf or .doc, will require a short teaser summary, and select the appropriate REAL rating.
-
Try using the questionnaire here: https://www.realgoodai.org/real-rating. The focus of this competition is human writing so a level of 2 or lower is required for submission.
-
AIs don't "think" at all. They are just regurgitating the most probable series of words like they are filling out a madlib. Maybe this (oversimplified) description will help illuminate things a little:
User inputs a series of words, this is called the context window
These words are then changed (embedded) into numbers. This is not quite right, but
These numbers are taken to a pretrained model and put in. You can think of this model as a very fancy version of filing in the gaps between two values.
The computer gives its answer as a series of numbers “filled in” from the gaps in between the data at the right place.
These are converted back into text that the user can interpret as language.
This is interpreted as "thinking" only because it is literally trained to sound like human text.
-
We are looking for unique stories that highlight real issues that we have with AI to educate the public to think while enjoying your art. There are plenty of depictions like this in sci-fi, and we think they are hugely inaccurate so it is best to avoid them if you would like to be selected.
The “learning to love” trope is scientifically incorrect because actually AIs already will use really emotionally loaded words and actions because they are trained to reflect how a human would respond. This is not the emotionless picture of calculating lovely logical, but instead will behave biased, flawed, and emotionally as a human does. Note that although it will have this appearance, there is no actual emotional effect, but instead is trying to emulate what a human may do.
As for sentience, this is also not scientifically feasible as AIs are just recognizing patterns rather than learning and demonstrating understanding. For the record, this does not mean they are not useful, it just means that they are not and cannot be sentient.
There are many issues and possible paths to explore so we would advise you to avoid these!
-
No. AI checkers do not work so our system for identifying AI writing is through the REAL rating you supply. AI cannot hold a candle to real human writing so we are not afraid of selecting AI stories because they will not be good like yours will be!
-
Sure! However, the prize remains the same for one episode so you will need to decide how to split the reward if selected.