Weekend Gems 008

Situating Methods in the Magic of Big Data and Artificial Intelligence

‍

For those of you who are interested in research that highlights all that exists at the intersection of ethics, technology and society, danah boyd is a name you need to remember. She is currently doing work on questions related to bias in "big data" and artificial intelligence, how people negotiate privacy and publicity, media manipulation, and the 2020 census implementation of differential privacy.Generally, Dr. boyd is a tour de force and I absolutely LOVE the fact that she treats issues like Big Data and Online Drama Culture with the same seriousness.

I recently returned to a 2017 paper of Dr. boyd titled “Situating Methods in the Magic of Big Data and Artificial Intelligence”. In the paper, boyd rewinds AI’s road to justification: starting from what legitimized Big Data some decades earlier, she first openly questions the idea that Big Data should be offering a tool of legitimacy to whatever argument they are occasionally embellishing. According to boyd, by accepting Big Data as the source of the One Truth,

we are automatically and uncritically disregarding established models of accountability while simultaneously redefining notions of ‘authority’ and ‘objectivity’

(we had briefly touched upon this idea in the very first WG ever).

What happens with most emerging technologies, following the Silicon Valley dogma, is that entire socio-technical concepts are structured as results-driven and later on adjusted to a process-oriented narrative(then of course something goes completely south and twitter urges the occasional company to re-examine its values. Also lots of public apologies and self-reflections). Same happened for Big Data: for example, when comparing the 2014 and 2016 reports of the Obama White House, one is met by a striking difference in the way even the Presidency treated this new concept. In almost two years and after a plethora of activists urging users and companies to pay attention to algorithms of oppression and discrimination in-the-making, the narrative around Big Data coming from the White House went from purely optimistic to painting a much more concerning portrait about the hidden potential of data discrimination.

Which led to a radical re-evaluation of the entire structure of Big Data and its socio-political implications, turning it into a fully transparent structure with the well-being of the new netizen, its anonymity and privacy in mind.

Lol jk they re-branded it as AI.

Which is bad for all parties involved. For one, what the business world—and consequently pop culture—currently seem to understand as “AI” is miles away from what AI truly is (or might be). As Dr. boyd notes,

today AI is a carrier of assumptions about intelligence, and different AI models stand for different approaches on the role of logic and reason in abstractly represented models of the world.

Granted, there is a certain novelty in the way AI research and progress have unwillingly (?) undertaken this performative role. It is borderline unprecedented to be dealing with a technology that is clearly oscillating between being a palpable scientific breakthrough and a purely cultural phenomenon, having immense impact as a socio-technical force even before its scientific legitimization.

A good example of such treatment, as boyd writes in the paper, was the CBS Sunday Morning news segment (2015) covering the DARPA robotics challenge showcase. This very real and serious topic was touched upon with the director of the science fiction film, Ex Machina, as the main panelist. That is, as boyd says, the morning news show used the director of a popular science fiction movie as the main commentator for the culmination of the most prestigious federally funded robotics challenge.

Please, take a moment to process this. If you have a hard time doing so, think instead of watching the news covering the latest discovery in cancer treatment and having Dr. Ramoray as the expert in the panel.

Yet, by treating this very serious matter in such a lighthearted way and allowing for such a-synchronicity between public perception and technical feasibility, we are prone to missing the “different approaches on the role of logic and reason” part that boyd said earlier. Simply put, we are jeopardizing the actual progress that we could be collecively marking in the field through asking the right questions, while obscuring the underlying issues and biases that are very much hindering in the real processes behind the real AI—namely pattern recognition, machine learning, deep learning and neural networks.

There is a massive shift happening right now through extensive research on deep learning tools—namely, a shift from creating symbolic representations of human knowledge and procedural knowledge, to generating new information from pattern recognition and statistic analysis’ strategies. We need to start paying attention to the latter.

The way we are collectively experiencing the rise of AI is like riding down a road of spectacles, witnessing a war between the technically feasible and the socially desirable. Such course is not necessarily bad as the popular opinion has proven historically to be a necessary driving force for research (usually bringing in the $$$). However, if this field is consciously driving down the path towards dominating the collective imagination while steadily scaling up normative logics, then all matters should be taken into consideration. If you are knowingly taking up a role, you need to pay attention to what this role stands for. For example, if you are going to use AI for spectacle, you should also be in the position to answer why Watson, the ‘smart AI that won Jeopardy!’ had a male synthetic voice whereas Alexa, an AI destined to serve your household, is marketed as female.

What I am trying to say here is that if we are going to build the myth of the sorcerer’s stone, we might as well do it properly because there is a chance that all we will be left with at the end of the day is the myth and no stone.

Highlight

‍“the logics, techniques, and uses of these technologies can never be separated from their specific social perceptions and contexts of development and use.”

‍Extras‍

‍Gil Press, quoted by Boyd herself, has extensively written about the mythologies of Big Data as well as some prediction for AI (in a rather pop manner).

As cheesy as the entire season was, a good example of the above is how Nolan imagined and designed Rehoboam, a massive sphere that had access to the entire data horizon and weaved predictions of extreme accuracy about society, economy and politics. The entire season kind of sucks but, paradoxically, I think the scenario and the main premise is by far the best of the three seasons.

‍

‘There’s no reason for an architect to design a death chamber’

‍

If only I had a dollar every time I heard someone saying “since it’s happening, it might as well happen right”. This all-time classic broken-logic card is usually thrown on the table every time someone is engaging with what they can easily tell is massively wrong but want to do it anyway. I know it because I have done it myself: as architects — especially while still in college — it is very common to participate in competitions of questionable morals, hidden under the veil of “paper architecture” (aka architectural design not made to be built). A student is not generally expected to question the deeper moral logic of competitions that otherwise propel, promote and sustain gentrification and colonization practices like building a 5-star hotel in an ‘urban neighborhood’ or designing a luxury villa in a protected landscape like a beach, a remote island or whatever shoreline. Similarly, competitions as such are usually not called out.

True: in theory, a competition and a realized project are miles afar; the former will forever remain in some corner of the internet while the latter will stand tall and very much concrete (!) for at least the next four decades. However, I would encourage you to ask yourself: in the context of a competition, would you ever design a prison? Or, as Michael Kimmelman asks, would you ever design a death chamber?

Michael Kimmelman recently wrote a piece in New York Times, asking A.I.A. to censure architects who designed prisons and death chambers. The article is written as a response to AIA’s statement in solidarity with protesters in the wake of George Floyd’s killing. In his own words while addressing the American Institute of Architects:

“So, for starters, how about stop repeating that it’s OK by you for architects to design death chambers and solitary confinement cells in racially biased prisons that incarcerate and execute an overwhelmingly disproportionate percentage of African-Americans?”.

Apparently, way before this month’s events, ADPSR—an initiative led by Raphael Sperry, had actually filed a petition to censure architects who designed death chambers and solitary confinement facilities. The response Sperry got from AIA was on the same intellectual level as the broken-logic card: since it’s technically legal and inevitable, we might as well be the ones to do it.

What this answer does is first and foremost reveal that the architectural community and everyone who aligns ethically with this justification is completely incapable of understanding the social context in which they operate in (while bragging about being able to affect and propel forwards the society they are addressing). In a less naive note, agreeing with this is straight up accepting that architecture is nothing more than execution of imperatives outside of the very field, rendering years of theory and critical making/thinking a merely performative aspect of the practice. The response wrote:

“Architects can’t be held responsible for what occurs in the buildings they design, and besides, while he had never been hired to design a death chamber, he was sure the right architect could design a more humane one.”

Trying to understand how a death chamber can be seen as humane in any context whatsoever is impossible so I am not going to even try at this point.

What we all fail to grasp however, whichever world or level of this world we are coming from, is that even the smallest of gestures is a statement. Whether that is a paper, a competition submission, a piece of code, a github repo, your latest instagram post — always remember that things remain. Things remain where we leave them as the internet is still struggling with the ideas of oblivion and erasure. Things remain and populate, and their population is usually outside of our own agency.

For architects, this means that your work does not need to be made out of concrete to “remain”; having koozarch post your project drawings has equal weight and importance—both things remain and through their lingering, they affect.

So placing a thing at whichever little corner of the internet we are occupying is our own way of structuring, enhancing and participating in larger discussions, preserving or cancelling ideologies and worldviews in the process.

All in all, it is a matter of participation: choosing to participate is to align yourself with the ideals in question—to become an ideological proxy of whatever it is you are participating in.

‍

And if you think that it is a bit far-fetched for architecture students to participate in competitions about prisons, I’ll just leave this here.

Life after WWDC and PRIMER2020

This past Thursday I celebrated my 100th day in quarantine. To celebrate it I decided to engage in a small retrospection exercise: what was that I gained or missed during these past 3+ months and what would that mean for the days to come, seeing the covid19 cases still on the rise here in the US and the chances for a second lockdown becoming bigger everyday?

One thing that became apparent to me pretty soon is that, for the first time ever and paradoxically enough, I had the chance to attend many events. The combination of the wfh model, saving me almost 20 hours of commute time on a weekly basis, along with the fact that I became significantly more immersed in the online space, led to me attending two online events — webinars, talks, conferences and symposia — every week on average.

Primer2020 just ended (and it was a BLAST—congrats to the entire Speculative Futures community for pulling this one through!), the 139th Arch League Annual Meeting is coming up and of course the Unreal’s Fest Online is just around the corner. The list goes on: just a quick visit to the tech meme events page will reveal the abundance of all tech events that are happening during the summer. The best part of it is that, most of these events, while pre-covid19 required one to take out their credit card and pay a sometimes significant participation fee, are all of a sudden offered for free. I honestly would not be more grateful for the fact that today I have free access to almost everything that in the past required me to purchase a sometimes hefty ticket, pack an entire suitcase, book a hotel, hop on a plane, travel to the other side of the planet and attend a keynote speech while having an internal struggle with my jet lag (although sometimes this was extremely fun).

Still, as Lauren Goode notes, something is profoundly missing here: the hang. Being able to just aimlessly stroll around posters, randomly engage in a discussion while grabbing a coffee during the breaks, or take a great discussion about the closing keynote speech from the venue to a local bar and brainstorm over beers is something that the digital space has yet to properly figure out how to recreate. One might say that this is because the virtual conference model is still at its early stages — not because the model has not found a way to replicate our past social interactions but because we haven’t had the time and space yet to radically re-invent and absorb what ‘social interaction’ means from scratch.

Hey - it takes time.

Trying to document the changes this new model has brought in the community, Lauren points out that online-only events are opening up the gates to normally underrepresented people at conventional tech events. From live translation to narration services helping the blind community keep up on equal terms with what’s going on in the conference, there is a surfacing of new tools and media that enrich the digital space and create new means of participation.

“Microsoft took it a step further during its Build event, offering not just closed captioning but bringing in ASL translators and offering them as a picture-in-picture option during lab sessions.” We should have seen it coming: of course tech giants as such, if pushed, would start expanding the capabilities of an online-only event. After all they are not conference organizers, they are digital space architects and if anything else, keeping up with how online conferences are being designed and held right now offers a good glimpse of how the digital space is going to be structured tomorrow.

‍

PS: Even if ‘the sense of community’ is going through some trying times right now within the digital space, I still think that adopting virtual models will be more beneficial for any community in the long run. After all, there is always a way to design spaces for community and comraderie online but it takes a miracle (or, apparently, a global pandemic crisis) to waive a $2,400 ticket.

Bonus Round

Did you recently upload a selfie with a face mask on instagram and carefully tagged it as such? Well, good news: there is a chance your face ended up in this GitHub repo , used to train facial recognition algorithms in the age of covid19. Once you let that properly sink in, you can watch John Oliver’s latest episode on Facial Recognition here.

‍Also, I visited Twitch for the first time and five minutes in I ended up at AtomicKawaii’s listening to her reading Harry Potter and the Scorcerer’s Stone. That was way more fun than I thought. I still don’t think that Twitch is my cup-of-tea per se as I am more positive towards more structured content with designed pacing but I never knew I would enjoy someone reading me a book on Twitch that much. ⭐️

‍