Extended Intelligences II
Last updated
Last updated
To start the journey into Extended Intelligences II class, as a continuation from the first term introduced during two weeks in the first term of the program, we were approached with a quick talk to remind ourselves about AI its implications, possibilities, and debates around the field and how people are approaching it along its evolution. However, different from the first edition, the professors, MDEF students already graduated, focused on showing more practical possibilities for using AI in real projects and works that have been applied in the market and exhibited in the artistic field, as a good real perception of real applications.
The presentation by Pietro Rustici and Christian Ernest showed us projects developed during the MDEF program during their time and after graduation, such as the use of artificial intelligence to create a gallery of portraits that collected news during the last Italy elections, a website that composes news from the future, and so a collective work reframing the idea of iconic design projects developed with AI. Through those ideas, they showed a little of what they achieved as a physical exhibition with the portraits in the Dutch Design Week and a physical translation process of the AI objects into physical artifacts that got invited to participate in exhibitions around Barcelona.
"It's moment of peak. Companies are spending money into new AI capabilities because no one wants to be behind and become a Nokia of the future."
As a continuous experiment from their journey within professional expertise in the area, the professors showed us a website they've developed in the last years with an interface that was designed on purpose for the user not to see the images captured by the camera. This funny website with an alien from Metal Slug on the loading screen was based on the stable diffusion model and it gave the user the possibility to write a prompt change the image by creating an alternative reality and vision of the picture without the user seeing the raw image, as a tool that we should use to create a collection of five images with a topic and concept in common made in groups.
By going to the "wild" and exploring the possibilities we could take advantage of, our group had some ideas of abstracting some elements and turning them into something else, but still with the curiosity to understand if the tool would change the environment or context according to the prompt since we used ourselves as models for the images. During the exploration, we sat on benches as cowboys, trying to mimic other animals eating plants, hugging metal abstract structures, and photographing pigeons to see replace the objects and animals interacting with fictional creatures and sci-fi scenarios.
As a result of our journey, we achieved some curious and strange pictures that related to reality on some different levels according to the prompt we wrote and how specific or open we were to the tool:
"An image of a wild dino human creature trying to feed itself with leaves in a forest."
My prompt has full detail so no need to add more. DO NOT add any detail, just use it AS-IS. Use the following information as the base details to generate the image: an image of a wild dino human creature trying to feed itself with leaves in a forest Scene description: The image depicts a person crouched down on the ground, seemingly examining a bush or plant. They are facing towards the greenery, leaning forward as if inspecting something closely. The person is dressed in a cream-colored jacket with black and blue stripes and is wearing light blue jeans. The environment appears to be an outdoor area with a mix of grass and bare patches of soil, indicating a garden or park setting. In the background, there are modern buildings with large windows, hinting at an urban area. Several leaves are scattered on the ground nearby, and some branches from the plant appear to be broken or fallen. The overall mood of the scene seems casual and curious, with the individual focused on the natural surroundings.
Generate an image of a wild dino-human creature attempting to feed on leaves in a forest. This being is crouched down by a bush, appearing to be examining the foliage intently. This dino-human creature is dressed in a cream-colored jacket adorned with black and blue stripes, paired with light blue jeans. The environment surrounding the creature is a blend of grassy and barren patches, suggesting a park in the middle of an urban area, hinted by modern buildings with large windows in the background. Scattered nearby are fallen leaves, with certain branches from the surrounding plants appearing snapped or fallen. The ambiance of the image emits a casual and curious atmosphere, signifying the creature's engrossed observation of its environment.
"Pidgeon dinosaur walking on a park."
My prompt has full detail so no need to add more. DO NOT add any detail, just use it AS-IS. Use the following information as the base details to generate the image: an image of a wild dino human creature trying to feed itself with leaves in a forest Scene description: The image depicts a person crouched down on the ground, seemingly examining a bush or plant. They are facing towards the greenery, leaning forward as if inspecting something closely. The person is dressed in a cream-colored jacket with black and blue stripes and is wearing light blue jeans. The environment appears to be an outdoor area with a mix of grass and bare patches of soil, indicating a garden or park setting. In the background, there are modern buildings with large windows, hinting at an urban area. Several leaves are scattered on the ground nearby, and some branches from the plant appear to be broken or fallen. The overall mood of the scene seems casual and curious, with the individual focused on the natural surroundings.
Generate an image of a wild dino-human creature attempting to feed on leaves in a forest. This being is crouched down by a bush, appearing to be examining the foliage intently. This dino-human creature is dressed in a cream-colored jacket adorned with black and blue stripes, paired with light blue jeans. The environment surrounding the creature is a blend of grassy and barren patches, suggesting a park in the middle of an urban area, hinted by modern buildings with large windows in the background. Scattered nearby are fallen leaves, with certain branches from the surrounding plants appearing snapped or fallen. The ambiance of the image emits a casual and curious atmosphere, signifying the creature's engrossed observation of its environment.
The initial project proposal aimed the ambition to create a communication language between different AI-integrated Arduino controllers with models for each cell that could sense their "siblings" and start a conversation using the sense of sounds, by hearing and speaking. However considering the complexity of the idea, during the first hours in class the group focused on developing one single cell using a microphone and a buzzer built within the Barduino controller, something that unfortunately didn't happen considering the low potential of the sensor that wasn't responding to any kind of audio interference, as claps, screams, voices and so on.
One extra sensor that was supposed to be used by the first idea, was the ultrasonic sensor, a specific electronic component that would measure the proximity between distances to trigger the first sound made by the Arduino, so the conversation would start initially by triggering their location and no radically making noises, what could interfere with the logic interaction. Even with a working sensor, integrating a microphone that wasn't responding to our actions wasn't showing us the possibility of continuing the project and delivering something tangible to the program, which made the group though more input and output possibilities as cameras with LEDs.
Besides the complications during the process, the group tried to reduce the elements and insisted a bit more on the microphone by verifying its installation through the Arduino board or even the code, something that unfortunately consumed a big amount of time to accomplish. With no results from the microphones and only the speakers during the two days of the project, the connections between our Arduino and AI kept presenting complications through different computers with different errors and outputs in an incredible domino effect where things were functioning at all, demotivating the group and so the ambition to achieve a project.
Within the final hours to deliver and present the idea, we changed the code with the help of the professors to achieve a simple mechanism using a potentiometer to collect that from human interaction and send it to the Open AI model used by their shared code to be analyzed and projects different outputs through the Barduino possibilities, as the RGB LED and the integrated buzzer. As an intriguing set of events, the code worked only on the professor's computer a mysterious problem considering we had the same software and libraries.
To conclude, during the project development, we even tried to use different AI models that could be integrated into Barduino without needing to use Wi-Fi, such as TensorFlow, but the libraries kept responding badly to the code made using Chat-GPT. As a consequence, the project had big ambitions and objectives but it truly needed more days and organization to understand the logic of the swarm intelligence and how to replicate it.
Unlike the initial part of this class in the first term, the second step of extended intelligences tried not only to remember some concepts presented months ago but also introduced new kinds of tools and language models that could be incorporated and translated into circuits and electronics interventions in the physical world. As a student, it was good to see previous alumni from the same course working for companies or studios with a good presentation of their transition from MDEF to the market, as a solid reality of the paths that could happen in discovering yourself through the program and then implementing in the real world.
Besides the good base and introduction from the professors, it's undeniable to recognize the low amount of time that is needed to ideate, develop, and prototype an idea, considering we had two weeks in the first term to develop only digital models, but it's impossible to deny the difficulty of the project chosen by our group to create during the two practical days in class. Within a complicated mish-mash of changes and adaptions between codes found, create, and formulated by some artificial intelligences, the project taught us to simplify and extract the essence of our idea.
As a biodesigner, this class made me reflect on how to mimic biological behaviors through AI and how they could be translated into the research I've been applying during the course so far, considering my ambition to not only replicate nature in its physical but mental for through the principles of computation. Trying to achieve swarm intelligence it's quite a complicated challenge, however by looking at some projects made by MIT and Harvard and through the errors plus technical mistakes made during the assemble process I could understand and identify better the points that should be improved to achieve an interaction/relation between autonomous machines.
After one week from the class, I've been trying to cluster the mimicking types of warm intelligence considering the use of AI models implemented into microcontrollers as Arduino's and which levels of difficulty each one demonstrates considering the amount of code and electronic parts necessary to make it feasible so our group could accomplish this objective since we still wanted to make the idea tangible. For me, developing these ideas could make me open new possibilities focused on biohybrid robots when not necessarily AI can give autonomy to the individual robots, but maybe analyze the environment as an observer and guide the whole herb of individuals to accomplish a specific test.
This theoretical scheme can clarify the goals of a project when not only developing one specific individual organism, but a mental model that can be replicated numerous times and how they behave individually or not between each other by creating their communicating system by sensing the environment and avoiding the unnecessary. At last, this reflection truly relates to my explorations trying to understand forms of consciousness and how they are applied to certain individuals, something that I could take more advantage of, but something to improve in the next Micro Challenge week, where professors will be there helping to comprehend the tools.
