Our online morgenbooster: Data & Technology as Design Material
The presentation will answer: What are the ethical and cultural implications of using data and technology as design material?
2. Per Jackson
Lead Creative Technologist
pcj@1508.dk
Online
Morgenbooster
Fredrik Silfver
CTO & Co-Founder at Above
fredrik@above.se
Anton Stonor
CTO & Partner at StrĂžmlin
ast@stromlin.dk
Data &
Technology
as Design
Material
3. AI - from pattern
analysis to
creativity and
design
23. John Maeda:
âDesign in Tech Report 2016â
https://designintech.report/wp-
content/uploads/2018/11/designintech2016_small.pdf âą
Classical Design Design Thinking Computational Design
Design of objects we use in the
physical world
Collaborate and innovate using
ideation methods
Creative activity that involves
computer technology and the
network
DESIGN BUSINESS TECHNOLOGY
39. Classical Design Design Thinking Computational Design
Design of objects we use in the
physical world
Collaborate and innovate using
ideation methods
Creative activity that involves
computer technology and the
network
DESIGN BUSINESS TECHNOLOGY
40. AI experts â Future plumber jobs?
SPECIALISTS GENERALISTS
AI Specialist
ML Scientist/Engineer
Deep Learning Specialist
NLP specialist
Data Scientist/Engineer/Analyst
AI Research Scientist
Data Annotation Expert
âŠ
Experience/Design Engineer
Design/Creative Technologist
Digital Designer
Data Visualisation Developer
Data Storyteller
...
42. What needs to happen?
Bring AI to the masses!
Democratisation of the
latest technologies
43.
44.
45. What does the industry need to do?
Bring digital transformation
& data into the core business
Bring developers into the
innovation process from day 1
Think data from the get-go
(Data/AI Driven Design)
1
3
2
Hinweis der Redaktion
I want to start off by talking a little bit about a term called âComputational Designâ. In 2016 John Maeda defined three types of design that influences the industry today.
Classical Design which deals in the design of physical objects like chairs, cars, etc
Design Thinking which pertains to how organizations learn how to collaborate and innovate using ideation methods
Computational Design which he defines as a âCreative activity that involves computer technology and the networkâ
And certainly weâve seen an increase lately in the third category - especially when looking at AI, Machine Learning and data use in design.
It seems everyone is collecting data these days and are trying to find new ways of incorporating technology in to our lives. And some of these are very successful - some of these are not so great...
One of the less successful ones was the âJuicero Pressâ. A $700 Wi-Fi connected juicer that automatically blends different kinds of juice flavors to meet your preference.
Turns out you could just as well mix the flavors yourself and achieve a better result.
Also, recently a new Nirvana song came out!
A Toronto-based non-profit organisation called Over The Bridge, wanted to raise awareness of mental health issues in the music industry. And so, they let an AI create a song based on the material of Nirvana and letting a Nirvana cover-band record the track.
Although a decent agenda, I think this also raises some questions on the consequences for the music industry and also the ethics about releasing music by essentially dead people.
Weâve also seen similar things going on in the movie industry, where dead actors are resurrected in new films.
But what does it do to the industry, when we suddenly have the ability to make ourselves obsolete - to be taken over by algorithms? I kind of see this scary scenario where everything will become copies - derived from what was once the pinnacle of human creativity.
But there are also great and really interesting things happening!
The Icelandic singer Björk has created a generative score of choir music for a hotel lobby in New York, that uses Microsoft AI to respond to changes in the sky above the hotel.
So using a rooftop camera, it creates a unique musical composition for the lobby and is played 24 hours a day.
I think this is a really creative way of leveraging AI technology for the physical space.
So talking to an AI - in this case the GPT-3 from OpenAI - is yet to become indistinguishable from talking to a human being.
But, weâve reached the point where itâs hard for us to detect if a composed text comes from a person or a computer.
This certainly has some implications for content creators, copy writers, journalists, etc. Will it just make their jobs easier, or will it end up replacing some of them?
And what about ourselves? Will we as coders be replaced in the future?
Weâre definitely seeing some progress in the world of No-code tools, and we are seeing the first attempts to let Neural Networks convert images to HTML code.
I think we are not quite there yet, but it might just be a matter of computing power!
In the narrow space of online surveys, Google Forms now uses AI to parse the texts from your questions to determine which type of answer will fit your question the best.
And this is really where I see a lot of potential in terms of Machine Learning. Especially within the realm of interface design, there are a lot of places where we could sort of sprinkle a little bit of Machine Learning on to our digital services to make them just a little bit better for the users.
And it seems to me that computers are really good at solving very specific tasks within a confined space. So when it comes to something like NLP or text recognition, there will be a fair amount of mistakes being made.
And so, we have to be careful not to expect right and wrong answers from Machine Learning, but think more in terms of probabilities. Mistakes will happen, so we as designers need to present our Machine Learning results in a way that is designed for errors - and perhaps say âthis might be a good option for you at this pointâ.
So can you tame AI?
Everyone is collecting data like there is no tomorrow! But to what end?
Facebook and Google are harvesting data on a massive scale and using AI for ad targeting your profile. And mostly it works pretty good - some would say too good.
But thereâs a flipside when we start trusting the machine too much or when we donât pay attention to potential biases hidden within our training models.
The National Gallery of Denmark has been harvesting a lot of data⊠Not about people, but about art!
Although not as complex as human profiles, artworks can be quite tricky especially when looking at some of the more abstract pieces.
Last year 1508 collaborated with SMK on this project call SMK Open, and while diving into this immense pool of data, we came up with a couple of interesting ideas - and essentially, we were using this data as a design material.
Not by using AI or Machine Learning, but by visualisation.
So each artwork has quite a lot of potential data attached to it.
And one data point was a list of locations where the artwork had been on exhibition.
So one of our designers came up with this idea of showing a map displaying the journey that each specific piece had been travelling through time.
I think this is a really fun way of telling a story about artworks.
We also realised that most artworks have a creation date.
And so we came up with this timeline where you would be able to see the distribution of artworks through history - also a really interesting way to look at their collection I think.
This feature didnât reach the final site unfortunately - but we imagined putting some of historyâs major events on this timeline to give the collection some context.
Right now we are also working with Finansforbundet - a Danish labor union for the financial sector.
And here we are creating a couple of services for the members to give them an overview of their work tasks and competencies. And in the process we are collecting data from their inputs in order to provide them with some useful ways to improve themselves.
Based on these data points, our AI Relation Engine returns a list of events, courses and articles that hopefully will help the users to amplify their skills and improve on their work situation.
So collecting data can serve a decent purpose, and itâs our job to find these opportunities within the space of Computational Design.
And speaking of jobs - the rise of AI have also seen the birth of a lot of new job titles out there.
Will AI Experts become the future âplummerâ job? I donât knowâŠ
But here are some of the job titles companies are seeking at the moment.
Thereâs a sort of division between Specialists and Generalists, where Specialists are more into the hard core development of custom AI infrastructure and services, while Generalists are more concerned with prototyping and coming up with new ideas on how to leverage AI and designing data visualizations.
As more and more things are getting connected, more data is being generated to power future customer experiences and business optimization.
So what needs to happen in order for us to keep expanding this train for the future we are on?
We need to bring AI to the masses!
And we need a democratisation of the latest technologies.
Whatâs the most important problem the AI community should be working on 2021?
This guy seems to hit an important point - âExplainable AIâ.
Fortunately these guys have stepped up to the challenge.
Although these âAI As A Serviceâ platforms can be quite a steep learning curve, we are seeing improvements every day, and the base entry level for engaging in these technologies are dropping fast.
So what do we in the industry need to do?
Well I think we need to:
Bring digital transformation & data into the core business
Think data from the get-go
Bring developers into the innovation process from day 1