Architecture, design and AI:
Emerging visions
WORDS Davina Jackson - excerpt from her recent book data cities
Architecture and design
are being transformed
by AI—artificial and
augmented intelligence.
Since 1936, when Alan Turing first theorised
an a-machine (automatic), our world has been
meshed with ‘intelligent’ devices. They process
information and perform complex tasks that
either cannot be achieved by humans or where
most people, relying only on natural intelligence
(NI), simply are not effective. The implications
for architecture and design education seem
immense and imminent.
Probably the most sophisticated and familiar
examples of AI are the mission-critical computer
systems which fly us around the world. Before
1914, when Lawrence Sperry demonstrated the
world’s first autopilot mechanism (using four
gyroscopes to stabilise a plane in flight), pioneer
aviators frequently crashed. Today’s aeroplane
control systems, when programmed with a flight
plan, automate the take-off, ascent, level flying,
descent and landing phases of flights. Also, they
constantly detect the positioning and behaviour
of aircraft in space: adjusting altitude, latitude,
longitude, pitch, roll and yaw. Similar electronic
systems now automatically supervise many
essential operations of buildings which never
leave the ground.
Autopilot systems control passenger drones—
flying taxis—which herald the huge new urban
planning challenge of how to manage cities like
vast airports. The Ehang 184, made in China,
flies with eight propellers on its four arms and
a host of sensors streaming real-time data, and
includes ‘failsafe’ backup systems in case of
emergencies. Flights are monitored by squads
of technicians watching giant screens in remote
control rooms. The only task to be performed
by a ‘pilot’ is to key the destination on a
smartphone app before automated take-off.
Drone taxis—coming soon from Uber, etc.—
revivify H.G. Wells’ Victorian flying machine
stories, which must have inspired the Wright
Brothers before their first flight in 1903.
Arrays of sensors also feed performance data
to America’s Cup sailors and cyclors flying
hydrofoil catamarans; to managers of public
surveillance, transport and emergency services
operations, and to technicians maintaining
‘smart’ city buildings and precincts.
Architects now can be conveyed to their
terrestrial sites in self-driving Teslas. But back
in their studios – seriously, could machines
really auto-generate building designs and city
plans? Not independently (or not yet) from
the mammalian architects and engineers who
program the algorithms and supervise/select
from the outlines of forms that they generate.
Yet there are some plausible progenitors
of creative appliances toiling in future ateliers
of design. For example, French inventor
Patrick Tresset has built a troupe of robo-
sketch artists whose electro-mechanical arms
rapidly scribble portraits via their cameras
and facial recognition software. Each has a
‘personal’ drawing style, which Tresset has
programmed—like architects guide their
drafting staff to follow ‘house’ drawing styles
and specifications. The robots, all named Paul,
are best compared after they have clustered
around a single sitter, scrawling jets of ink
across paper on old-school wooden desks,
just like humans in life-drawing classes.
architecture & Design /
PEOPlE
/ Jul-sep 2019
18
ADQ3_018_021_AI Architecture_V1.indd 18 26/7/19 4:48 pm