SoylentNews is people
SoylentNews
from the that's-just-what-the-bot-wants-you-to-believe! dept.
New AI fake text generator may be too dangerous to release, say creators
The creators of a revolutionary AI system that can write news stories and works of fiction – dubbed "deepfakes for text" – have taken the unusual step of not releasing their research publicly, for fear of potential misuse.
OpenAI, an nonprofit research company backed by Elon Musk, says its new AI model, called GPT2 is so good and the risk of malicious use so high that it is breaking from its normal practice of releasing the full research to the public in order to allow more time to discuss the ramifications of the technological breakthrough.
At its core, GPT2 is a text generator. The AI system is fed text, anything from a few words to a whole page, and asked to write the next few sentences based on its predictions of what should come next. The system is pushing the boundaries of what was thought possible, both in terms of the quality of the output, and the wide variety of potential uses.
When used to simply generate new text, GPT2 is capable of writing plausible passages that match what it is given in both style and subject. It rarely shows any of the quirks that mark out previous AI systems, such as forgetting what it is writing about midway through a paragraph, or mangling the syntax of long sentences.
More like ClosedAI or OpenAIEEEEEE.
Related: OpenAI 'Universe' Platform Provides Means to Link Image Recognition Vehicular AI Agents Into GTA 5
The OpenAI Dota 2 Bots Defeated a Team of Former Pros
(Score: 2) by JoeMerchant on Friday February 15 2019, @04:14PM (2 children)
Yeah, that's where my prototype code was at in 1984 - 35 years later, with a million times the computing power, and the life's work of 100s of developers to build on, they're doing better now - maybe not perfect, "true AI" always seems 5 years away, but the old Lincoln nugget: "you can't fool all of the people all of the time" portion of people-time who are not fooled by these things is continually shrinking.
🌻🌻 [google.com]
(Score: 2) by acid andy on Friday February 15 2019, @05:19PM (1 child)
I guess, as is so often the case, it all depends on the training data. For fiction we're probably close to most plausible common human events having already been described by someone, somewhere, so with a large enough data set, I suppose you could make a convincing work of fiction by pattern matching all the source material. I doubt the AI will know whether the events it's describing obey the laws of physics but I guess plenty of human fiction violates those too. This reminds me a bit of Searle's Chinese Room where abstract symbols are being manipulated without needing to be fully understood.
If a cat has kittens, does a rat have rittens, a bat bittens and a mat mittens?
(Score: 2) by JoeMerchant on Friday February 15 2019, @07:06PM
With the recent explosion in entertainment production for Netflix et. al., I feel like a lot of the shows I try to watch are mashed together by a lame AI pulling elements from previous successful shows. In a way, that's what the money backing productions does: looks for "guaranteed, maximized" returns by putting out entertainment with known popular (i.e. money making) components.
🌻🌻 [google.com]