Agency & Creativity
A friend of mine shared a previous issue of my newsletter, “Can Machines Be Creative?” while discussing using Artificial Intelligence, namely Generative AI, for App Development.
To recap, there are three types of creativity discussed in that issue. They are:
Combinational Creativity
Exploratory Creativity
Transformational Creativity
The last issue was that Machine Creativity can do combinational creativity, help out on exploratory creativity but very likely cannot do transformational creativity.
After a very insightful discussion with Eric Sandosham, I kind of missed out on something in my discussion, or I did not manage to get to the very basic principle of whether Machines can be Creative.
All Generative AI need prompts to generate the different prototypes for Combinational and Exploratory Creativity!
Let’s take a step into understanding what prompts are. Prompts are made up of words or statements that is used to pass instructions to GenAI to generate the desired output. While they are made up of words, these words have to come from humans.
Humans have to base on the desired output, generate a set of words that influence the “environment”, in this case, the Generative AI model. Thinking further, Creativity requires agency to come up with new and original ideas. This is true from Combinational Creativity to Transformational Creativity! From another perspective, unless machines start to gain agency, we cannot conclude that GenAI is creative but rather a tool to materialise new ideas, and humans, with their agency, mesh these ideas together.
Creativity is coming up with ideas. These ideas can be through birth (original), or TAKING two or more ideas and combining them. The birth and combination of ideas have to come from the human. Humans through prompts as input into Generative AI, which comes up with the output to better assist the viability of the output.
To conclude, agency is needed for creativity to take place, and agency remains steadfast in the domain of humans rather than machines…until there is a way for us to reliably test if AI has agency. Will they?
What are your thoughts on this? Share them below! :)
I very much enjoyed the discussion, which presented many perspectives. If you do, consider reaching out to me on LinkedIn.
Consider supporting my work! You can make a “book” donation and drop me some wisdom! :)
Below are a few past issues you might be interested in.
