The document discusses the use of the Craiyon AI, formerly known as DALL-E, to create visual art from text prompts, emphasizing its unique capability to generate imaginative images. It reflects on the interaction between human prompts and AI-generated outcomes, noting both the limitations and the charm of machine-emulated art. The author also explores notions of creativity, machine learning, and the evolving nature of artistic interpretations in the digital age.

1. Co-Creating
Common Art with
the CrAIyon AI
Shalin Hai-Jew
August 2022

2. Preface
• The alcohol ink drip playing series continues but with less of a sense
of urgency. (For those who need a fix, here are some AI-created
images emulating alcohol ink drip playing art. I did not include one
that had a suggestive part to the image. Remember that the AI learns
from people.)
2

3. 3

4. 4

5. 5

6. 6

7. 7

8. 8

9. 9

10. 10

11. 11

12. 12

13. 13

14. 14

15. 15

16. Preface (cont.)
• During a recent 3-day academic conference, one of the keynoters
mentioned CrAIyon (https://www.craiyon.com/), formerly DALL-E
(after Salvador Dali). This AI is now CrAIyon as a mini DALL-E, which
learns from a mass training set of images and uses text as an input to
create visual art. The fine print says that the free images can only be
used for teaching and learning, in terms of the non-commercial
access.
• When CrAIyon thinks of Dali…it thinks of visual gists.
16

17. 17

18. 18

19. 19

20. Preface (cont.)
• Indeed, the idea is to use various textual prompts that are complex
and fairly unique to trigger the AI to create cool visuals. This
slideshow will show some of the images evoked although I am leery
about naming the artists. (The AI picks some of the most notable
features of the respective artists, and it almost looks like a send-up of
their work sometimes. That is not my intention, of course.)
• There are times when I ask for something “photorealistic” and am still
shocked at how real-world the contents are.
20

21. Preface (cont.)
• I have a delight when the AI depicts something that does not exist in
the world but could if one were in a machine’s imagination.
• A machine “imagination” does not exist, and nor does machine “sentience,”
even though both have been part of sci-fi for decades.
• The AI seems to be cobbling elements…and it seems to maybe work based off
of an overarching simplification of shapes for coherence and then fills in the
rest.
• In terms of credit, the human mostly only things of the text as the
seeding concept…and then maybe engages in light digital
cropping…and in some cases more. In my case, I did not do much
more. I just am charmed and intrigued.
21

22. A Machine and its own Rules
A Jigsaw Puzzle Series
22

23. The Jigsaw Puzzle Series
• I like how the jigsaw puzzle images might suggest a machine that
squishes puzzle pieces, so they fit.
• It looks like the machine may be changing the rules of the jigsaw
puzzle game.
• It looks like the machine is dreaming about twisting the cardboard to
make a piece fit.
• Or perhaps I am reading into the visuals.
23

24. 24

25. 25

26. 26

27. 27

28. 28

29. 29

30. Machine-Emulated Art
30

31. Three Modern Artists
• I picked two well-known artists from the 20th century with well-
known signature styles and a large body of work each.
• Then I added a third who is still living and working in a different
medium.
31

32. 32

33. 33

34. 34

35. 35

36. 36

37. 37

38. 38

39. 39

40. 40

41. 41

42. 42

43. 43

44. 44

45. 45

46. 46

47. 47

48. 48

49. 49

50. 50

51. 51

52. 52

53. 53

54. 54

55. Machine Motion
55

56. And Motion?
• What is motion in a world without physics?
• What does an AI see second hand (based on human tagging…machine
tagging)?
56

57. 57

58. 58

59. 59

60. 60

61. 61

62. 62

63. 63

64. Mixed Outcomes
64

65. 65

66. 66

67. 67

68. 68

69. 69

70. 70

71. 71

72. 72

73. 73

74. 74

75. Conclusion
• These are all an idea expressed textually, an AI trained on human-
created imagery, and outcomes in pixels.
75