Machina ex Machina: 2014 by pistolwish

Machina ex Machina: 2014 by pistolwish

https://vimeo.com/361937943

Pistolwish Selections from Machina ex Machina. 2018-2019 AI Generated Video - 60-second excerpts sampled from 2014 and 2019 www.pistolwish.com Up to a decade ago, news was limited to an industry defined by reporters and stories filtered through human editors. Then, factually prepared and delivered through daily papers, or read to you in produced segments on televised broadcasts. Since then, the news has been democratized in real time with camera phones turning observers everywhere into reporters. The excessive amount of unedited user-generated content along with information coming from traditional and emerging sources has become too much to consume without assistance. People, have turned to AI to interpret and recommend what they see and hear in order to be informed. For this installation, a random set of recent news articles was interpreted and matched to videos depicting the same meaning. The process was rerun on videos from 5 years prior as a comparison . The resulting video demonstrates how AI programs can reach conclusions from unedited user-generated content. For instance, between 2014 and 2019, 63% of the matches featured strong violence associated with the same topic. This is evident through observation across each topic and visible in the footage samples. AI programs are making choices for us everyday - what street to drive, which songs to play, who you should date, the jobs offered to you, and even which stranger’s cars you’ll jump into. AI programs are useful tools for addressing the problem of information overload. However, we must consider it’s biases and weaknesses whenever presented its interpretations. About the project: Machina ex Machina (‘MeM’’) is an ongoing project exploring how AI programs interpret each other. In its current invocation the MeM, AI was taught to watch videos and read the news to understand what they’re about based upon its content. For this project a random selection was made from my AI-curated news feed from three distinct sections (World, US, and Entertainment as appearing in Google News on or about Sept 1, 2019). MeM read the articles and determined its meaning, then selected closely-related videos it watched. The process was run again, but trained on data from 5 years prior with random samples from each group added to the comparison videos. Future installations will revisit the same topics at different time intervals, as well as explore other sources of AI-interpretations and content.

Uploaded 2019-09-24T00:08:00.000Z Aerials and Satellites Wakefield