Research Presentation

The Algorithmic Layer

Mapping machine intelligence as a system in the landscape. How do recommendation algorithms create feedback loops that reshape our physical surroundings?

Kimberly Blacutt / American Society for Cybernetics, DC Arts Center Gallery, 2024 / 45 minutes, Presentation + Discussion

Every day, more than a billion people ask an algorithm where to go. What happens to the places it sends them? What happens to the places it doesn't?

The Premise

In landscape architecture, we map ecological systems to understand how they interact: hydrology, topography, soils, vegetation, circulation patterns. Each layer reveals forces that shape the land.

What if we could map machine intelligence the same way? Navigation algorithms from platforms like Google Maps and Waze now direct billions of daily decisions. These recommendations create feedback loops between digital systems and physical space.

This research proposes a new layer for analysis: the algorithmic layer.

The Layer Cake Model

Systems within a landscape

As a landscape architect, I'm used to thinking about relationships between different ecological systems within a geographic setting. I've typically represented different types of social, biological, and physical systems as layers.

Layer details

Each layer contains specific data, examples of which are circulation networks, vegetation patterns, or hydrology.

Systems overlay

Examining and overlaying these different layers helps us find points of synergy and contention — which we can then use to make better informed design decisions.

Site analysis outcome

For example, in choosing the best location for a new development we overlay layers that reveal transportation routes, ecologically valuable areas, and watershed drainage patterns.

Layer details — circulation, ecology, hydrology

Systems overlay — layers combined

Site analysis outcome — systems informed decision

Recently, I've been thinking that there is a hugely impactful layer missing in the traditional layer cake model — one that I think we should consider in these types of site analysis exercises.

A Missing Layer?

What might we call this layer that actively shapes how we experience and act upon the landscape?

The Algorithmic Layer

This is a live information feedback layer that produces recommendations for people's actions. I call it the algorithmic layer.

A missing layer? — question mark layer above the stack

The algorithmic layer revealed

So, let's examine this algorithmic layer. What kind of algorithms am I talking about?

I'm talking about any digital platform that uses an algorithm to recommend something to you — because those recommendations lead to actions.

The Pattern

Algorithm → Recommendation → Action

The algorithmic layer follows a consistent pattern: algorithms process data, produce recommendations, and those recommendations drive human actions and decisions.

Yelp + TripAdvisor

Review platforms optimize for ranking. Their algorithm recommends restaurants — and we eat there. The algorithm shapes where we go and what we experience.

Yelp and TripAdvisor algorithm diagram

Instagram + TikTok

Social media platforms optimize for engagement. Their algorithm recommends content — and we view, share, and engage with it. The same pattern, different domain.

Instagram and TikTok algorithm diagram

Google Maps + Waze

Navigation platforms default to optimizing for lowest travel time. Their algorithm recommends routes — and we travel on them. The same three-step pattern — and I argue that this recommendation has direct, physical consequences on the landscape.

Google Maps and Waze algorithm diagram

Algorithm → Recommendation → Action flow diagram

Person eating at a restaurant recommended by an algorithm

Person watching cat videos recommended by a social media algorithm

Person following a navigation algorithm's fastest route

Reflection Question 1

What are the consequences of algorithmic optimization? Are the values being optimized reflective of societal values?

Let's take a closer look at Google Maps as a case study, because its recommendations have direct, physical consequences on the landscape. More than a billion people use Google Maps every month.

The Dataset

What feeds the algorithm?

Google Maps relies on a massive, complex dataset. It aggregates over 1,000 authoritative sources, satellite and street-level imagery, and data from more than one billion monthly users.

The feedback loop

Interactive maps create an instantaneous feedback loop between the user and the data. The reading of the map modifies the map. Interacting with the data itself gets quantified and becomes data.

The massive dataset that feeds the Google Maps algorithm

The feedback loop between user behavior and algorithmic data

Reflection Question 2

Is there important information that is missing from or misrepresented within algorithmic datasets? What language can we use to name, reveal, or correct the data?

How might algorithms be provoking changes in the physical landscape? I think that algorithmic recommendations do shape our everyday actions and those actions have the potential to change the landscape.

To give us all more concrete images to ponder, here are a few real-world stories.

Real-World Evidence

When Algorithms Reshape the Landscape

Isometric illustration of a car approaching a collapsed bridge while GPS says 'You are on the fastest route'

The Guardian

Family sues Google after Maps allegedly directed father off collapsed US bridge

Tech company faces negligence lawsuit after Philip Paxson died from driving off a North Carolina bridge destroyed years ago

In 2022, a man died after following Google Maps directions onto a bridge that had collapsed nine years earlier. The platform's data had not been updated. This raises questions about standards of care: when we trust an algorithm, who is responsible for its accuracy?

Read full article

Isometric illustration of artist pulling wagon of smartphones across a bridge, creating virtual traffic jam

Hyperallergic

Berlin Artist Walks the Streets With 99 Smartphones, Prompting Google Maps Traffic Jams

Simon Weckert's conceptual Google Maps Hacks leads us to question the power of maps as a daily tool and control mechanism.

Artist Simon Weckert pulled a wagon containing 99 smartphones across a bridge over the Spree river in Berlin, creating a virtual traffic jam in Google Maps. The platform interpreted the clustered signals as congestion, rerouting real drivers to other bridges. Art as algorithmic intervention.

Read full article

Isometric illustration of residential street overwhelmed with cars while residents watch, with 'Not a Highway' sign

Los Angeles Magazine

Waze Hijacked L.A. in the Name of Convenience. Can Anyone Put the Genie Back in the Bottle?

Traffic apps turned the city's neighborhoods into "shortcuts." Now furious residents are attempting to take them back, street by street.

Los Angeles residents found their quiet hillside streets overwhelmed with through-traffic as Waze routed drivers through residential neighborhoods. Residents fought back — reporting fake accidents, petitioning for speed bumps and one-way conversions. This is one of the clearest examples of how the algorithmic layer ends up reshaping the physical layers.

Read full article

Reflection Question 3

What is the conversation that we have with the algorithm? What are the consequences of that conversation?

The Threshold

These stories reveal a conversation we may not realize we're having.

Every time we follow a recommended route, we generate data that reinforces the algorithm's model. Compliance becomes confirmation. The map learns from our behavior and adjusts.

But in some cases, the feedback loop does not stop at the screen.

When navigation algorithms reroute thousands of cars through residential streets, neighbors respond not by changing their phones but by changing the landscape itself. Speed bumps are installed. Streets are narrowed. Signs appear. One-way conversions are proposed. In some cases, residents attempt to pass new local legislation.

At this point, the algorithmic layer has crossed a threshold. A digital recommendation has triggered physical infrastructure changes, political action, and new rules of movement.

These interventions are material attempts to push back, slow down, or confuse a system optimized for efficiency rather than livability. The neighborhood becomes an interface. The street becomes a site of negotiation.

The algorithmic layer does not simply sit on top of the physical world. It reshapes it and is reshaped in return.

Reflection Question 4

How can we reveal and describe the ways algorithms already shape our landscapes? And how might this understanding allow us to design, or resist, the futures they produce?

Want to talk about this? Email me at kimberly@blacutt.com