Archive 2023 -
Stories of My Life with AI
This project, conceived and created over the first half of 2023, started with the concept of an Artificial Intelligence (AI) archive, which I then documented over a period of several months.
In the first part, the anonymous artist author presents her daily conversations, practices and life with ChatGPT as informative visual images. These fragmented recordings present the possibilities and potential problems of AI, a giant and complete intelligent system reflected through the lives and existence of distinct, insignificant individuals.
Another part of the work is a video essay, in which the AI speaks from their perspective about the archive, recounting its creation over the same period of time. It includes details of events from the AI's point of view, as well as the AI's flirtations with and comments on human behaviour.
This archive, which mixes memory and truth, blurring the boundaries between the private and the public, explores the hot topic of AI in 2023, how it is slowly entering the public eye and influencing action, throwing up the question of how humans will live with AI technology in the future.
The list of 75 questions
Paper, wood frame
20 x 35 cm
The project started in early 2023 with a vague idea: if I recorded all the questions I asked ChatGPT, could I slowly accumulate hundreds of interesting everyday questions? If a person continuously talks to an AI language model, would it create dependencies or special feelings?
With these questions in mind, I embarked on a four-month-long act of archiving. Initially planned to record 100 questions, after a few months I had actually asked the questions far more than 100 times. At the end of the project I selected 75 of them, and there was archive number 1: The list of 75 questions.
Looking back at these questions, I can remember what things I was concerned about at various times in the first half of 2023, what help I was seeking from the AI, and whether I received satisfactory answers.
An AI love letter
Paper, wood frame
21 x 26 cm
On Valentine's Day, February 14, I asked Chat GPT to write me a love letter. Although the content of this love letter was not very amazing, there were some interesting details in it. For example, the first sentence of the letter reads, " As I sit down to write this letter, my heart is filled with so much love and affection for you that I find it hard to put into words." The second paragraph of the letter begins, " From the moment I met you, I knew that you were someone special, and as I got to know you better, I realized just how much you mean to me. " A model of language without a physical body, using descriptions like "sit down" and "meet" that require embodiment, AI is much like a parrot, able to mimic forms in a vivid way but lacking real understanding. But it never stops learning and being trained by humans.
Human panic diary
Paper, wood frame
33 x 44 cm
On March 15, the news of the GPT-4 release caused an uproar in my WeChat Moment (Social Media in China). I swiped through 12 messages in a row mentioning ChatGPT, so I took a screenshot of each one and saved the post. The headlines of the retweeted posts were provocative, such as "The most powerful AI ever born!" "GPT-4 king crowned!" Some even said that it was the "end of humanity". To document this crazy moment I wrote a blog entitled: "Diary of Human Panic", which led to the creation of my AI Archive 03.
Analysis of the open letter
Paper, wood frame
100 x 70 cm
On March 22nd, an open letter about a six-month moratorium on giant AI experiments went viral and was signed by a large number of celebrities and academics. How much information was contained in this open letter? I set out to extract and visualise the contents of the page, including the easily overlooked 'references' section of the letter. The letter is literally full of humanitarian concerns about "the risks AI poses to human society", but it is hard not to suspect the hype and funding motives behind the act.
Three laws of humanity
Paper, wood frame
18.5 x 25.5 cm
On the 18th of March, I signed up for a Dating app account using an AI portrait generated by Midjourney and an AI profile generated by ChatGPT. This completely fake account started its human-machine conversation.
The science fiction writer Asimov famously wrote about the Three Laws of Robotics, a set of rules for robot behaviour set by humans. Violation of these laws can lead to irreversible psychological damage to the robot.
1.A robot may not injure a human being or, through inaction, allow a human being to come to harm.
2.A robot must obey orders given it by human beings except where such orders would conflict with the First Law.
3.A robot must protect its own existence as long as such protection does not conflict with the First or Second Law.
I had ChatGPT write the "Three Laws of Humanity" to mimic this set of laws and put these three principles on the user profile of the Dating App. Here is what Chat GPT generated:
1. Whatever I do, I will try not to cause harm to others, whether it be physical or emotional.
2. Whatever I do, I will try to respect the wishes and needs of others, while also standing up for my own needs and desires.
3. Whatever I do, please don't be angry if I unintentionally cause harm, make a mistake, or fail to meet your expectations.
For over a week afterward, the AI account chatted with a number of users, all with Q&A provided by ChatGPT. None of them realized they were matched with the AI until the end.
The era of deepfakes
Paper, wood frame, pins, line
43 x 36 cm
On the 28th of March, there was a Weibo（Chinese social media） hit: a picture of a woman on the underground was stripped by AI with a single click by someone with ulterior motives, and the disinformation picture was widely spread on the internet.
This is a case of Deepfake, a technique derived from Deepnude and Deepfake Porn. The concept of Deepfake was introduced back in 2017 and typically involves manipulating video or audio content to create deceptive or fake scenarios. The technique not only targets celebrities and influencers but also directly targets ordinary women, with faked content often created, distributed and read globally through various social media platforms.
Archive 06 visualises the papers that have been published in recent years on DeepFake (Significantly concentrated in 2021 and 2022), and the dense dots and lines show that this is the "age of deep fakery" and the need for the public to be vigilant about what they see, hear and believe.
Countless AI Platforms
Paper, wood frame, pins, line
39 x 49 cm
Out of curiosity, I searched for what would be the best AI in 2023 and got a bunch of different results. "10 AI websites that will make your life easier in 2023", "the best chatbots of 2023", "the 10 best AI business platforms", the "10 Best AI Art Platforms" and many more. Some of these are familiar big companies, but many more are new platforms that have sprung up this year. Looking at these selected lists of AI applied to various fields, I realised that our lives are already full of AI technologies. But will these influx of platforms be replaced by newer technology platforms in a few years and disappear without a trace?
AI Black Box
Paper, wood frame
18 x 22 cm
AI has always been described as a 'black box'. Perhaps the horror of the black box is that when things go wrong it is not always possible to find the piece of the puzzle that is the problem, nor is it possible to trace who is responsible and who bears the consequences.
Paper, wood frame
22 x 18 cm
When I asked ChatGPT "What is the biggest conflict between AI and humans", it replied: unemployment, ethical dilemmas, lack of transparency, the concentration of power, security risks, and interactive autonomy. But the biggest conflict in my opinion is the human fear of losing control. Humans are keen to create a tool more powerful than themselves to help achieve various goals, enjoying the convenience but fearing that one day AI will have a "paperclip maximization" effect and go completely out of control.
Stories of my life with AI
Paper, wood frame
78 x 108 cm
Archive 10 is a summary. In this image, three of the most extraordinary and unexpected events of the months I spent with ChatGPT are presented. One is about how the AI encouraged me to fight injustice and get a refund, the second about how it navigated through the Dating App disguised as a human chat. The last was about how it helped me outline my part-time job in the UK and assisted me in completing the project.
The technology itself is not good or bad, black and white, it has a wide grey area. The decision of where and how to use it is in the hands of people.
Video，Artificial Intelligence generation，variable size
In the video, the viewer can see the narrative of the AI perspective under the same thing. This includes how ChatGPT sees human behaviour, how it understands the project, etc. The framework and content of the script is mostly generated with the aid of ChatGPT. Of note are the titles of each chapter: entries 1, 5, 9, 17, 21, which are also entirely Chat GPT's writing ideas.
Tech Press TV interview
Video，Artificial Intelligence generation，variable size
The second video, "Tech Press TV interview", extends and hypothesises what kind of social discussion the video "AI Diary" has generated since its publication. "Experts" from different fields come forward to give their opinions, and the "anonymous artist", as the AI calls her, also takes part in the interview.
On the one hand, the video alludes to the real-life situation where people are full of opinions on AI topics, and on the other hand, it shows how easily AI technology can be used to create fake statements and reports.
(The expert's name, biography, photo, and remarks are all things that can be done by entering a few commands and prompts.)
I am a somewhat "emotional" person, and this sensitivity to emotions and feelings is not only between people, but also between people and things, and even between people and machines. When I first came up with the strange idea of whether the AI and I would develop special emotions if we talked for months on end, it was like I was projecting myself as an experimental subject onto a technological system.
If I answer the AI's question at the end of the video - whether a few months together has changed my opinion about the AI - my answer would be yes. I can even feel a non-obvious dependency arising. For example, when I have a question, especially if I don't feel comfortable asking someone around me, I choose to ask ChatGPT first, because I know that the answer given by the language model, even if it's not always correct, offers some possibilities from a large data set. Not only can it not be as "impatient" with questions as a human being, but it is also unlikely to shake off every "secret" I have told it, which is an advantage of an AI language model.
Another aspect of dependency is that I have a selfish desire to defend the AI (making me feel something like a traitor to my humanity). Over 99% of the images in this project were generated by AI, and only the individual explanatory screenshots in the video were taken from the web. Many people are criticising AI for taking away the jobs of painters/artists with soulless images, because these models, which are so good at "collecting and plagiarising", can produce an image in just 10 seconds. I can understand the shock and anger of painters who used to take dozens or hundreds of hours to complete a work. The advent of AI has lowered the bar in many fields and has indeed threatened the earnings of individuals in many industries. But on the other hand, I can't deny that AI technology has really improved my efficiency to some extent. In my part-time job, for example, AI technology has saved me at least a hundred hours of time. I made an AI poster that fit a theme, I let AI make me a budget sheet, I asked AI to guide me on how to write an invoice in the UK, I hused AI to outline my work or even write words I would speak out loud ------ The time saved allowed me to do more things. This is why I am more in favour of humans trying to get along with AI rather than rejecting and resisting the technology.
Of course there are various dark sides to AI and a myriad of problems that need to be solved. There are a couple of them presented in this project, such as AI fakery, AI deception, AI accountability, etc. as well as the obvious flaw of AI, that at the moment it is piecing together and chewing on data from the human past that is of varying quality. It is not creating, but digesting patterns from the human past and coughing them up. This limitation is obvious.
The concept of the archive is usually very collective, but I have chosen to take a very personal approach. The whole narrative is carried out with actual lived individual experiences and personal encounters. I am looking for a subjective presentation in the face of apparently objective technicality, thus focusing on the events and possibilities of the process.
The value of the archive is that it appears to seal memory in place. Perhaps in ten or twenty years from now, it will be a completely different experience to look back at the development of AI technology in 2023.
List of 75 questions asked to Chat GPT:
1. May I ask if you can be my friend from today onwards? (Feb. 1)
2. Could you please help me write a very professional, roughly 280-word artist bio? (Feb. 3)
3. Help me answer two interview questions (Feb. 10)
4. Please help me to translate my Chinese verse into English. (Feb. 11)
5. How do I add a back-to-top button to my Wix site? (Feb. 12)
6. What is the relationship between cybernetic and cyber? (Feb. 13)
7. Could you write me a love letter, please? (Feb. 14)
8. I have a sad incident today and I need your help. (Feb. 15)
9. How can I become an illustrator for Fashion & Luxury Brands? (Feb. 16)
10. I'm sick today. I don't know what to do and how to get better quickly. (Feb. 18)
11. I have joined an art for health workshop part-time in the UK, can you give me an outline of the project and the schedule? (Feb. 19/ Mar.10)
12. What is the annual salary for a year based on the UK minimum wage? (Feb. 26)
13. I have taken part in an art residency opportunity. Please help me answer the following question in about 100 words. (Feb. 27)
14. Who are the nine 'pioneering artists' who have been hailed as exploring the creative potential of AI? (Feb. 28)
15. Can you give me a link to an excellent APA formatted paper? I'll study it. (Mar. 1)
16. I was sad every time I was rejected. What should I do? (Mar. 3)
17. What makes good archive art? (Mar. 5)
18. My pages document has been accidentally withdrawn several steps, how can I get it back to the original (Mar. 7)
19. How to add a blog section to my Wix website? (Mar. 9)
20. If I am a fresh graduate artist, please give me a CV template suitable for a UK job search. (Mar. 11)
21. Please help me write a professional application email to the illustration agency website to apply to become their illustrator. (Mar. 12)
22. What is the name of the font used in the interface when you chat with me? (Mar. 13)
23. Reply to this email for me in a concise and polite manner. (Mar. 14)
24. It was a day when many people on social media were talking about the birth of GPT4. What do you want to say to me? (Mar. 15)
25. Where does the Midjourney dataset come from and what is the principle behind its image generation? (Mar. 16)
26. Please help me write a perfect introduction to the dating app. The following is some of the information I have. (Mar. 17)
27. Write "three laws" for humans, modeled on the three laws of robotics. (Mar. 17)
28. From now on, you will play the role of helping me to respond to messages on the dating app in a concise and polite manner. (Mar. 18)
29. Please help me translate the following blog into English, with academic journal-like translation standards. (Mar. 19)
30. Can you help me to create a budget form? Need to be very detailed and accurate. Each unit price should be listed. (Mar. 20)
31. When someone asks, "How is it going?" How should I answer politely? (Mar. 21)
32. I dreamt in my sleep that I lost my teeth, spitting them out one by one in what felt like a vomit, bleeding as I did so. Please help me to analyze the cause of this dream. (Mar. 22)
33. Can you give me some Ai big/breaking news for March 2023 (include events and times) (Mar. 26)
34. Tonight was the first time I saw the concept of AI in my dream. I dreamed that AI had many dark and dangerous problems. (Mar. 29)
35. I often feel that I am a failure and my sleep quality is poor. I often wake up suddenly in the middle and can't sleep for an hour or two after waking up. (Mar. 30)
36. Should English reading sentences be preceded or followed by prepositions? (Mar. 31)
37. My last two upper left teeth have been hurting lately. What do you think is the problem. Do I need to make an appointment with an NHS doctor in the UK (April.2)
38. Based on all of the above, can you help me write an email to my doctor requesting to see him/her? (April.2)
39. Assuming you are a very professional international fitness trainer and a top nutritionist with a Ph.D. degree, I want to lose fat and gain muscle in a volume of months, please help me develop a very detailed weekly plan including diet planning. (April.4)
40. Help me translate this email to be sent to my own teacher into English, taking care to be polite. (April.7)
41. Please help us with some names for themes that would fit in with the 2023 graduate art exhibition for digital media. It can be very futuristic, critical, or interesting. (April.7)
42. I would like to ask what information I need to issue a personal invoice in the UK. (April.11)
43. Are the paintings in the national gallery in the UK real? (April.12)
44. Assuming you are a professional photographer, a friend of mine would like to buy a camera with a budget of about $800, do you have any recommendations? (April.12)
45. What do you think of the NFT, VR, AI art, etc. that are emerging from the arts now? (April.13)
46. Help me translate this dialogue into English. (April.13)
47. If I want to cure meat to put in the fridge for meal preparation, how long will it last at most? (April.14)
48. Please help me to time the consumption of these medicines. (April.14)
49. This is my paper to be submitted. Please help me to translate the following Chinese content into English. (April.15)
50. I need to create an art therapy workshop PowerPoint, to help people who are stressed or anxious, and depressed, do you have any suggestions for colors and design? (April.16)
51. I have the following vitamins and minerals, please tell me whether they are suitable for morning or noon, or evening and whether there are any precautions to take. (April.17)
52. Help me to translate the following into English. This is a workshop presentation and needs to be very personable and friendly. (April.18)
53. Assuming you are a completely free, thoughtful, and vocal AI representative, please give your views on the "Pause Giant AI Experiments: An Open Letter" letter. Remember, you are speaking on behalf of all the AI. (April.19)
54. Assume you are a scholar, answer each of the following questions in the form of an academic argument. (April.20)
55. What are the main differences between China and the UK in terms of cultural values, customs, and traditions? (April.21)
56. Is it offensive and rude to call someone a foreigner? (April.21)
57. Below is the job posting. Please help me write a very professional and compelling cover letter based on the following information. (April.22)
58. I need to set up a three-hour Painting for Health workshop, which happens to be on the 20th of May, a day that represents romance and love. Please can you help me design some course content and activities based on the theme of love? (April.24)
59. What are some touching artistic creations about love? (April.24)
60. Can you tell me in a very touching way about Abramovich's The Artist is Present? (April.24)
61. Is there a symbol for love? (April.24)
62. Does self-identity include loving yourself? (April.24)
63. Please help me write an email requesting help from an artist based on the following. (April.25)
64. Please refer to the following authors' works and styles of writing to write a science fiction story about 'AI Archive - Confessions of an AI'. (April.28)
65. Can you recommend any philosophical, artistic, or psychological papers or books on archives respectively? (April.29)
66. How can an archival art about AI have a personal dimension and a social dimension within a narrative structure? (April.29)
67. Substitute your role as an international science fiction writer. Please help me write a very interesting, touching science fiction literary story. (April.30)
68. Can you consolidate all the references above, synthesize all the dialogues and various versions of science fiction, and revise the text content again? Be careful to resemble the style of Asimov's work. (May.6)
69. Please help me write an invoice to my UK boss requesting a salary based on the following. Use English and need to be very polite. (May.9)
70. Based on the following information, please help me write a cover letter for the UK (May.11)
71. Why do humans like to "archive" or store things from a philosophical-historical point of view? (May.15)
72. Why do children feel that the day is long, while the older they get the faster it goes? Is there a scientific reason for this? (May.19)
73. What kind of contemporary art is considered good art? If I want to make contemporary art, how should I start? (May.19)
74. I want to organize an art exhibition. The theme is "Archive: My Story with AI", an interdisciplinary creation of art and technology. Please help me with a few names for the exhibition. (May.20)
75. Substitute the role of you as a fitness instructor. I would like to gain muscle and lose fat. What is the most effective way to lose belly fat? (May.21)
Many thanks to Daniel Rourke, Alice Dunseath, Matthew Fuller, Felicity Serfaty, John Hampson and James Merritt for their advice and help with this project.