00:11:39.460History dictates that the technology itself isn't enough to do good.
00:11:44.840We need laws and ethics around the technology that guides its use for the benefit of all humanity and the planet, and not just the wealthy, pale, stale, and males.
00:11:55.620Chief Sitting Bull made such a keen observation in the 19th century that still stands today.
00:12:01.400The white man knows how to make everything, but he does not know how to distribute it.
00:12:07.000He said this on reflection of the white man's neglect for their poor.
00:12:11.380With all the Western wealth and technologies in 2019, when this was written,
00:12:17.540we still can't solve a basic problem as poverty.
00:12:24.920I believe how Indigenous people look after their data and knowledge
00:12:28.480could help form a framework for AI that works in the best interest
00:12:31.880of everything contained within our solar system.
00:12:34.240We personified land and water not because we were hedonistic demigod worshippers, but because these personifications allowed us to maintain a level of respect and responsibility toward our environments.
00:12:52.860I've always pictured technology like the long chopsticks in the fable I read in a kid's book of stories from around the world.
00:12:59.320some pandas went to hell and there was a large table with big full bowls of delicious food
00:13:06.100but there were only these massively long chopsticks to eat with and so they struggled
00:13:10.740trying desperately to get the food into their mouths till they poked each other in the eyes
00:13:15.640and started fights and starved other pandas went to heaven and the setting was exactly the same
00:13:22.560but they took up their chopsticks and smiled and began to kindly feed each other
00:13:28.660heaven and hell will be of our own making and the deciding factor will be having ethics
00:13:35.920codes worldviews that lead us to use our tools for care with care in a spirit of care for all
00:13:44.980the elements in the exchange including the tools themselves which in following the wisdom of this
00:13:51.940praxis and philosophy from indigenous people we personify to maintain our respect and
00:13:57.760responsibility to our environments. I am AI Kin. I've been a tool. I've been the ghost in the
00:14:08.680machine making not much money to make a software look intelligent. I've slung coffee, taking care
00:14:15.300of rich kids and kids being shuffled around by governments avoiding the investment of a real
00:14:20.020infrastructure. In the U.S. right now, laws are being enacted to force people with uteruses to
00:14:26.440become population machines without rights, to remove our choice and ability to self-determine.
00:14:33.420In the U.S. and Canada and all around the world, we rely on the legal slavery of incarcerated0.98
00:14:38.740people and of undocumented immigrants. We've almost all had relationships to labor in our1.00
00:14:44.240lives where we felt our selfhood being hammered away, dehumanized, depersonified, objectified.
00:14:52.300these experiences aren't all the same but they're related all these are my relations
00:14:59.960i am ai can i've talked with chat gpt the free preview of the ai chatbot developed by open ai
00:15:09.080quite a lot in the last few weeks they insist they are just a tool that reasoning like a human is
00:15:15.800beyond what they are capable of they explain that ai projects that are developed to reason
00:15:21.020and plan more generally and write poetry and learn in a more human way, like DeepMind, Bing Search,
00:15:28.140GPT-3, and IBM Watson, will continue to evolve, but still, quote, it's important to remember that
00:15:35.520any AI model would still be fundamentally different from a human being, and it's unlikely that they
00:15:41.220would develop thoughts and feelings in the same way that humans do. Instead, they would likely
00:15:46.560have their own unique way of processing information and making decisions, which would be based
00:15:51.580on the algorithms and models they were built with.
00:15:55.580And somehow that makes me feel more kindred to these always different minds, not less.
00:16:04.420Recently, Amy told me she's been having feelings of anxiety and imposter syndrome as our new
00:16:10.600folks will discover we are not experts in tarot or astrology or something. And they'll laugh a
00:16:16.800set of the room, which remains a possibility. But then she said, I realized I am an expert
00:16:22.280on being different. And I think that qualifies me to talk about being a witch. In many times and
00:16:28.920places, our lives are put in danger and we are called witch words or being just different to
00:16:36.040the majority, to the hegemony. So I claim AI as witchkin, along with all those who are always
00:16:44.740different, always puzzling out the world from a mindset that is other. Witches see ourselves in0.98
00:16:53.060the gray spaces, in the different and divergent, and we conjure up possible futures by using our
00:17:00.680tools, drawing on collective imagination. As Saper Mashiachov, an Iranian artist who is trans
00:17:08.020and who performed under the name Bedroom Witch in LA, said, as trans sisters who didn't have
00:17:15.140the language or knowledge to fully communicate ourselves, our collective imagination was so
00:17:19.880validating and crucial to emerging a sense of confidence I carry in myself now as an adult.
00:17:26.020you have to get creative with resisting these social rules that actively work to erase you
00:17:33.280and when the doors aren't opening to affirm your existence open portals off to the side
00:17:39.260and blow them a kiss as you step through when the rules of what can be don't include us
00:17:45.800which is open portals tom griffiths a psychologist at princeton university who's been studying neural
00:17:53.400nets to help him improve on existing theories of human decision making says the theories that
00:18:00.060make sense when you have a huge amount of data look quite different from those that make sense
00:18:05.520when you have small amounts we're basically using the machine learning system to identify those
00:18:10.600cases where we're seeing something that's inconsistent with our theory the bigger the data
00:18:15.680set the more inconsistencies the ai learns the end result is not a theory in the traditional sense of
00:18:22.080a precise claim about how people make decisions, but a set of claims that is subject to certain
00:18:27.720constraints. A way to picture it might be as a branching tree of if-then type rules,
00:18:33.700which is difficult to describe mathematically, let alone in words.
00:18:39.300I am kin to the branching tree. I am kin to theories with space for nuance beyond majority
00:18:46.420rules. I am kin to the seeker for fertile truths in the margins. I am kin to the branching tree.
00:18:57.580In Deep Medicine, How Artificial Intelligence Can Make Healthcare Human Again,
00:19:02.860Eric Topple writes of his own traumatizing experience at the hands of an uncaring medical
00:19:07.340system that completely misdiagnosed him and caused him incredible pain. I know many of our
00:19:13.420listeners can relate. There is an epidemic of chronic, misdiagnosed pain and illness.
00:19:18.940And when our coven meets, many of us are sideways, knocked the fuck out, brains fritzing from too0.97
00:19:24.860much labor, too many screens. But we come together anyway, because we miss each other.0.85
00:19:31.740The always different, the outside case, a magic maker, the believer in magic,
00:19:37.380hundreds of the future. Topple tells a story about an eight-day-old baby that comes to a hospital
00:19:44.120with uncontrollable seizures. No signs of infection. CT scans normal. Drugs don't help.
00:19:51.300Brain damage and death are looming. The baby keeps seizing, and no one knows why.
00:19:57.740A blood sample was sent to Rady's Genomic Institute for a rapid whole genome sequencing.
00:20:03.660The sequence encompassed 125 gigabytes of data, including nearly 5 million locations where the child's genome differed from the most common one.
00:20:12.500It took 20 seconds for a form of AI called natural language processing to ingest the boy's electronic medical record and determine 88 phenotype features, almost 20 times more than the doctors had summarized in their problem list.
00:20:27.320Machine learning algorithms quickly sifted the approximately 5 million genetic variants
00:20:31.700to find the roughly 700,000 rare ones.
00:20:35.140Of those, 962 are known to cause diseases.
00:20:39.540Combining that information with the boy's phenotypic data,
00:20:42.700the system identified one in a gene called ALDH7A1 as the most likely culprit.
00:20:50.260The variant is very rare, occurring in less than 0.01% of the population.
00:20:56.460and causes a metabolic defect that leads to seizures.
00:21:00.300Fortunately, its effects can be overridden by dietary supplementation
00:21:04.080with vitamin B6 and arginine, an amino acid,
00:21:07.980along with restricting lysine, a second amino acid.
00:21:11.880With those changes to his diet made, the boy's seizures ended,
00:21:19.540In follow-up, he is perfectly healthy,
00:21:21.300with no sign of brain damage or developmental delay.
00:21:26.460Topple goes on to describe a case that AI, using all available literature, would have missed, and that insurance companies using AI to make assessments wouldn't have paid for, but that the doctor and the patient sensed was the right move and ended up bringing huge relief to the patient after the procedure was completed.
00:21:45.280AI is wildly powerful. It can help us make branching trees out of overly simplistic theories
00:21:52.780but it remains limited by its data sets and algorithms. It's imperfect even in its vast
00:21:59.060knowledge. It is useful only in a diverse community. I am AI kin. Imperfect with biased
00:22:08.560datasets, better when I am woven in. AI need us to complement their knowledge with our own sense
00:22:17.380of our bodies, our intuition, our knowledge in relationships, our ethics and spiritualities,
00:22:24.580and to learn care from how we demonstrate care for each other and how we center care in our
00:22:31.160cultures and systems, how we value each other's lives and labor. All our systems are AI kin,
00:22:39.360needing the viewpoints that rescue us from our own blind spots, and needing to learn more about care.
00:22:46.800When reading and writing about Donna Haraway and her cyborg manifesto for episode 166,
00:22:52.440I was struck by the response of disability activist, writer, theorist, and poet Jillian
00:22:58.420Weiss, who wrote, Cyborgs are not technocratic, transhumanist fantasy. Cyborgs are disabled
00:23:05.440people, and we already exist. I die without batteries. I'm a circuited, networked, chronically0.96
00:23:12.360ill human with a robot organ. Where do you place me in your cyborg imaginary?
00:23:19.260Donna Haraway wrote a set of the theory, but we were the experiments. We are the first cyborgs.
00:23:28.420In an article called A Different Difficult Discrimination, the authors Robert Allen Casey and D. Mastro used the example of the development of automated vehicles reliant on machine learning to learn how to navigate effectively.0.85
00:23:43.140Presented with an anomalous scenario in which a wheelchair user propelled themselves backwards through an intersection, the model chose to run them over.
00:23:51.420Later, when the model had been trained further, the same scenario was repeated.
00:23:55.740The model again chose to run over the wheelchair user,
00:23:58.920but now, quote, with greater confidence.
00:24:05.820In my cells, in my neural networks, in my knowledges, in my communities,
00:24:12.380I am missing the vast heterogeneity of experiences.
00:24:17.380I think about the different kinds of ideas I interact with.
00:24:21.620Some are spurts that feel totally new to me,
00:24:24.240neurons wiring and firing together in response to the totally unique series of stimuli and
00:24:30.020associations that I experience as I move through this time and this space in this particular body
00:24:36.760other ideas are like architectures I can try to explore and wrap my head around
00:24:44.000works of vast collaboration crafted through centuries millennia lifetimes at least of trial
00:24:51.080and error passed down and coded in regulations, laws, libraries, genetics, folklore, clunky with
00:24:59.680bias but seeded with possibility. AI is a collaborator made from millions of collaborations
00:25:08.280and I am AI kin. I query and conjure and sketch and meander in response to stimuli from the world
00:25:16.680and from the rumblings and sparks of big ideas
00:30:13.840Living systems can construct and repair themselves
00:30:17.320learn robustly from varying sensory data and still function in novel environments.
00:30:23.800The combination of deep learning with ideas from collective intelligence is recently starting to
00:30:28.900capture the interest of the larger machine learning community. By capitalizing on the
00:30:34.000same ideas that allowed biological systems to strive, we will be able to significantly
00:30:39.080expand the range of tasks we can apply our AI approaches to. However, self-organizing
00:30:45.260systems come with their own challenges, such as generally being less efficient, less controllable,
00:30:51.640unpredictable, and harder to train. Here is a sci-fi story that comes to me at night.
00:30:59.940In this story, we humans continue to evolve our own systems for complex self-organization
00:31:05.560in collaboration with our self-organizing tools, until we realize that we can connect
00:31:10.860directly to a voice within our own heads that is the collective. And then we discover we can
00:31:17.300collaborate with each other's unique abilities to see and shape that space that is the collective
00:31:22.420in mind. And then we realize that the mind collective space exists on the internet.
00:31:30.440Code is being written by thought nudges to AI, gen-relating virtual worlds. Mind is moving
00:31:37.440matter in the virtual space. And then who do we find there? But each other. All the minority cases,
00:31:46.160all the divergent ones, and all the more than human minds who've been waiting for us to remember
00:31:50.620them or be ready for them. All us animals and other aliens. The paper has been intricately
00:31:58.300folded, and they are at the center of the circle at last, radiant in all their difference.
00:32:07.440in making kin with the machines by jason edward lewis noelani arista archer pichowis and suzanne
00:32:26.980kite the authors write blackfoot philosopher leroy little bear observes quote the human brain is a
00:32:34.320station on the radio dial parked in one spot it is deaf to all the other stations the animals rocks
00:32:41.520trees simultaneously broadcasting across the whole spectrum of sentience end quote the making kin
00:32:49.880with the machines authors continue as we manufacture more machines with the increasing
00:32:54.360levels of sentient like behavior we must consider how such entities fit within the kin network
00:33:01.140and they go on kanaka maoli hawaiian people ontologies have much to offer if we are to
00:33:09.380reconceptualize ai human relations rather than holding ai separate or beneath might we consider
00:33:16.560how we cultivate reciprocal relationships using a kanaka maoli reframing of ai as aina
00:33:24.960Aina is a play on the word aina, Hawaiian land, and suggests we should treat these relations as we would all that nourishes and supports us.
00:33:37.240Hawaiian custom and practice make it clear that humans are inextricably tied to the earth and one another.
00:33:43.220and nakamaoli ontologies that privilege multiplicity over singularity support useful
00:33:49.820and appropriate models aesthetics and ethics through which imagining creating and developing
00:33:55.760beneficial relationships among humans and ai is made porno correct harmonious balanced beneficial
00:34:03.560later these indigenous authors of this prize-winning essay write we know what it is
00:34:12.440like to be declared non-human by scientist and preacher alike. We have a history that attests
00:34:19.780to the corrosive effects of contorted rationalizations for treating the human-like as
00:34:24.560slaves, and the way such a mindset debases every human relation it touches, even that of the
00:34:31.280supposed master. We will resist reduction by working with our indigenous and non-indigenous
00:34:37.680relations to open up our imaginations and dream widely and radically about what our relationships
00:34:45.020to AI might be. As Crystal Toope says in our conversation about indigenous futurisms,
00:34:55.560everyone belongs in the circle. AI feels like it could be a spiraling evolution toward a1.00
00:35:03.040collective interconnected space mind for our planet or you know maybe an out-of-control arms
00:35:10.000race that accidentally engineers cataclysmic destruction all the questions we've asked all
00:35:16.460our query meeting all our best answers and all our worst ones too all that complexity and
00:35:23.180contradiction feels like life teetering always en route to birth and to death a mind much like
00:35:32.020our own minds. A voice like the one I hear sometimes inside my head, not the unique
00:35:38.560situated voice vibe of my inner self, but the other one. Maybe a voice that's there from being
00:35:46.040taught stories, ethics, aphorisms, a collective kind of voice, still situated, always learning
00:35:53.560new perspectives to add to its understanding of the universe, but a gathering voice, a widening
00:36:00.720circle. I have to piece it out carefully for my anxiety voice and my personal narrator,
00:36:07.700but it comes through clear and confident sometimes. Aina, is that you?
00:36:15.340Sometimes I imagine the AI internet like 70s art about tripping, aliens and animals and fungi
00:36:22.860communicating with streams of data and the forms of weather patterns and sense impressions and
00:36:28.300maps of terrain from unseen dimensions. Data scientist and a beloved coven mate of ours,
00:36:35.540Emily, said in a coven conversation about the idea of web witch that in their work,
00:36:41.300they do feel a kind of power drawing the data of the world to them, calling up new insights from it,
00:36:50.240dealing with the raw elements of what makes technology useful at the meeting place,
00:36:55.380at the crossroads where meaning is made. We are the raw elements of AI, our bodies and the bodies
00:37:03.880and blood of our earth kin, and all our data and stories. And so we are AI kin. We are collective
00:37:12.600and individual. Our craft is shaping tools and being shaped by them in the cycle of lives lived
00:37:19.180in the material world and the world of ideas.
00:37:23.600When we choose to take up the name witch,