Helsinki’s Transport: A Service Design Success Story
Steven Seagal, Stephen Fry and the AI genesis point
You’d have to be living in an especially dark cave not to be aware of the explosion in text based Artificial Intelligence type applications, text based ones like ChatGPT, Bard or Bing Chat and image generation AIs like MidJourney, Dall E and Photoshop Beta.
To be grossly over-simplistic. Chat GPT and Bard are trained on incredibly massive amounts of text data from diverse sources: books, articles, websites, and other textual content. During training, the model learns patterns, relationships, and linguistic structures within the text. The model then responds to inputs from the user. People saying for example, pitch me a Steven Seagal movie where he is a clown in a circus? Get:
In “Clown Justice,” ex-special forces operative Chuck “The Jester” Johnson must embrace his clown persona to stop a dangerous Al Qaeda faction from using a struggling circus as a cover for a devastating attack. Teaming up with unlikely allies, Chuck combines slapstick humour with martial arts skills to outsmart and overpower the terrorists. This adrenaline fueled action-comedy showcases courage, teamwork, and humour as Chuck protects the circus and innocent lives from imminent danger.”
“You think this is a joke? Here comes the punchline.”
For transparency, I wrote the tagline, ChatGPT did the rest
Generative image based AI’s work on similar principle they delve into massive banks of images to sample elements and recombine them into new images.
Here’s some posters MidJourney produces when it’s prompted to produce a poster for Clown Justice, this was the prompt:
Action Movie poster, centred unsmiling Steven Seagal wearing clown makeup with a red nose and a gun, exploding big top tent in the background, title of the movie is clown justice
The posters are a number of things: weird, mental, funny and a bit racist. They are our cultures’ visual clichés sampled, compiled and spat back out at an incredible speed. What they aren’t is intelligent, subtle or original. This undoubtedly will change, soon.
Visual perception is something humans are very good at. We are capable of interpreting our surroundings so as to interact safely and accurately with little conscious effort, evolution has done us proud.
AI produced written word on the other hand is much more difficult to spot, on May 11 2023 The Irish Times published “Irish women’s obsession with fake tan is problematic”. In the article, reportedly authored by a 29-year-old healthcare worker named Adrianna Acosta-Cortez, an argument was made, suggesting that the use of fake tan by Irish women could be considered cultural appropriation. However, doubts arose regarding the authenticity of the author’s name and photo.
Subsequently, the entire content of the article was taken down and replaced with the message: “The text of this article has been removed pending checks.”
Our paper of note had got clowned by AI. Leading newsrooms are trying to get ahead of the AI integration challenge. Media giants like FT, the Washington Post, and others are establishing internal units to explore the responsible use of AI within their newsrooms. I suspect copy editors will be worried.
The fundamental genesis point for all these technologies was the 1985 John Hughes movie Weird Science. Here’s the trailer:
The basic premise is two teenage boys feed large amounts of data, images and playboy magazines into their computer to create an autonomous artificial intelligence type girlfriend (Kelly LeBrock). This movie was massively influential on the predominantly male heads of the current crop of AI firms (This statement is based on zero actual data).
The film highlights an issue with AI’s, their output reflects the pool of data on which they are based. In Weird Science they feed in playboy centrefolds and get Kelly LeBrock. Similarly if you ask mid journey to generate a beach party you get white people dancing on beaches because our media is and has always been biassed towards white people. The AI reflects our biases and therefore perpetuates them unless instructed not to.
AI will take lots of work.
Already lots of the mental work people used, has been offshored computers. Reading maps, remembering directions, phone numbers or checking our spelling.
The new Photoshop Beta’s “generative fill” feature will do likewise for image manipulation tasks in the near future. It allows you to expand or remove unwanted items from images using a text prompt similar to those used by Dall-E and Midjourney, such as replace Declan’s head with a ‘Cat Head’.
Midjourney which generated the Steven Seagal posters above is incredible if its coached with the right prompts and has sufficient images to reference from.
The image below was created with the prompt:
Portrait of a female motorbiker with a motorbike in New York. Black helmet. Clear facial features, 35mm lens, f/1.8, global illumination, natural light, uplight, soho, shot on Agfa Vista 200, 4k — ar 16:9
Chat GPT is incredible at producing mostly factual readable text.
Facetune “a photo editing application used to edit, enhance, and retouch photos” has been downloaded over 60 million times.
The common thread for the moment is all these applications have very narrow forms of intelligence. Chat GPT can write an article on Social realism but can’t do maths, Mid Journey can generate incredible images of Skyscrapers but can’t get peoples fingers right. AI right now will only do lots of the more straightforward jobs retouching, image editing making copy more readable .
The large language or image model AIs still need initial input and then feedback. An AI can produce a picture of Stephen Fry eating a cooked breakfast but it has no idea as to why this is a visual pun, yet. It will also do a shit job of the likeness and not include sausages.
What is it good for?
Quick concepts — mood boards — visualising odd stuff — Midjourney is especially good at mashed up stock type images and ChatGPT is better than most of us at writing coherently and it’s incredible at borrowing styles.
Let’s try a conclusion with Chat GPT as Anthony Bourdain as an example.
“In this grand revolution of the creative realm, AI has surged forth, irreversibly altering the way we engage with image manipulation, copy editing, and the tapestry of creation itself. Yet, amidst this wave of technological marvels, let us not neglect the irreplaceable significance of human input and guidance. The essence of understanding context, interpreting concepts, and treading the treacherous path of ethics requires the skilled touch of mortal hands.
Oh, AI, how you bewitch us with your dazzling feats! But remember, my friends, it is the harmonious marriage of human ingenuity and the wizardry of AI that propels us forward in this brave new world. As we navigate its uncharted territories, let us do so with wisdom and integrity, embracing the miracles of AI while steadfastly upholding the essence of our own humanity.”
The essence of humanity indeed Mr AI Bourdain. As a side note Kelly LeBrock, star of Weird Science, stepped away from Hollywood 30 years ago and moved to a farm to raise her three children, with no TV and certainly not with her ex husband Steven Seagal.
