Podcast transcript: DigiLab Online review (2020/21 Semester 1)
This podcast is part of the UCIL Digital Society course from the University of Manchester running in 2020/21 semester 1. The story it relates to is hosted on Medium and can be found here.
In the podcast, Jade and Naman from the Library Student Team review the topic, discussing some of your thoughts and advice on DigiLab Online.
TRANSCRIPT
Hi, my name is Jade from the Library Student Team and I’m here with my Student Team colleague Naman. We very much enjoyed reading your contributions on DigiLab Online’, which was this week’s topic. Today, we will be discussing and elaborating on what you said in response to the interactive activities and questions on this topic.
By now, you have already tried out different online experiments that we have compiled. We have collected a lot of different insights on how your experiences are, and what you think about them. Thank you for sharing your perspectives with us, as they make the course more engaging and give you a chance to try out new technologies.
The first experiment that you have reacted with is ‘Quick, Draw!’, which is a neural network that guesses what you are drawing. In the comments, some of you said that the technology was “very quick and accurate, and it exhausts all possible options before getting to the right answer”. Most of you said that the network was able to get most (or all) of your drawings. You commented that “the drawings the technology missed may be because of your own drawing skills”. I’m no artist myself, but these technologies know that the people playing may have different skill levels and we urge you to think about: how does it affect people with different skill levels? Is there an optimal skill level where the technology can guess everything you draw?
Next, you played around with Interplay Mode, which has prototype interactive videos using AI to direct you to play a virtual piano. Many of you were disappointed with this, because it was “confusing to use” and “it assumed piano skills that you didn’t have”. Others found it disappointing for other reasons. You said, “A very underdeveloped and simplistic programme, especially in a developing world of Electroacoustic Music. There are so many other programs where the computer can be programmed to respond to music in a more sophisticated way, e.g. MaxMSP”. As you can see, varying familiarity with musical software resulted in different perspectives on this prototype. This begs the question: is there a one size fits all in AI? How do programmers accommodate presumed knowledge into their coding and software design?
Your next activity was GauGan, which uses Ai and segmentation to create completely unique images. In response to this, many of you found the dance movements “strange”, “bizzare”, and “all too similar to one another”. One of you had a very interesting insight, “This tool shows how AI can be used fundamentally as an archive. We can preserve our past ideas in order to re-use them and create new, in this case, dances. I think this shows potential for the use of AI as it means we could preserve ideas through generations. AI would serve as a living library of ideas and developments. This tool also emphasises how AI is not just for the experts but can be used by amateurs” We agree! This tool seems to get around the problem of varying skill sets proposed by Interplay Mode and Quick Draw. But what, if any, is the tradeoff cost for that?
Next, you played Semantris, which is a word association game based on a machine learning algorithm, based on billions of examples. The idea of this was to reduce time spent watching a screen, so the software used a ‘paper phone’. Many of you commented on the hypocrisy and irony of this. You said, “ it does not fulfill the primary purpose of the phone — to stay connected to others. How do I communicate with my peers and family through the paper phone?” and “I use my phone to talk to people and keep up to date with social media and I am not sure how a paper phone will do this.” These are all very valid comments and have pointed out that even the most well-intentioned software can have its limitations and gaps. From this, how can programmers and developers build on software like this to, in the end, reduce the amount of time people spend using said software?
You also tried ‘GiorgioCam’, which helps turn your photos into music. While some of you were very intrigued by “how it did it”, others were skeptical. You were concerned about how “the data from games are saved and if yes, how they will be used.”. Indeed, data privacy is incredibly important and we all need to think about how our data is used to help companies keep track of us and learn more about us. One of you thought about the long-term implications of software like this, “Although entertaining, this is not in itself one of the most beneficial uses of AI but shows that we must experiment to understand how we can use it to help in different sectors”. Indeed, the applications of AI are vast but we need to think about how it impacts people and how it can be used to make a positive difference.
Lastly, we asked you to try ‘Experiment with Google’ to find different AI and machine learning technologies and play around with them. The answers you gave us were diverse, interesting, and showed your comprehensive understanding of the subject. One of you, “found an app which stops us from scrolling for hours through our social media accounts. The further down you scroll, the more the screen starts to look like the sea, which means that suddenly fish are swimming around, the screen gets darker and at some point you have reached the bottom. This app can be very useful to keep us from scrolling through Instagram or Facebook for hours.” After the disappointments with the paper phone, it is nice to find an example of a software which is better at making people more aware of their screen time.
That’s all from us this week. Thank you for engaging with the topics and taking the time to write down wonderful, thoughtful comments. We enjoyed reading them very much and we hope to hear more from you next week!
