Mar 14, 2019
In his second column on Design and Artificial Intelligence Karel Golta wonders how designers can make usable AI-processed data.
What a wonderfully bizarre picture: Shenzhen, the tech mega-city in December - on the office floors and in the conference rooms people meet the industrious and hard-working colleagues wearing thick winter jackets. Are they freezing more than others or is it an energy problem? Not at all. The answer is much simpler: for the few weeks in the year when the temperature drops below 18 degrees, it is more efficient to have employees wear jackets than to equip buildings with heating technology.
The situation is quite different with the equipment around the topic of Artificial Intelligence. The photo of the AI-controlled Smart City Observatory in Longgang, Shenzhen will certainly not only impress NASA, but would also serve as the backdrop for a science-fiction blockbuster. This is where the gigantic amounts of data that are collected throughout the city and processed with the help of AI flow together.
Besides the coats in the office topic I constantly wondered: How can the more than 30 employees mentally grasp and contextually process the entire column, diagrams as well as other information graphics displayed on the enormous display? How useful is it to turn innumerable data streams into a multitude of cake and bar charts or heat maps? Even sudoku experienced employees should find it difficult to create the dependency to the right heatmap from the left upper middle diagram compared to the lower right bar graph. Seriously, is AI processing the flood of data or just maintaining it?
Shouldn't it be the power of AI to create value from data so we can make decisions faster, easier or more meaningful? Or is it just a design question? Is the central screen simply not designed to be user-friendly? Do designers need to learn to make better graphics in the age of AI? Do design and petabytes of data fit together at all?
AI does not work without data. But data has no value if it is not used. This is where design comes into play. Because designers create value by making something useful - more and more through AI.
The work order is thus: Create value X (output) for target group Y (user of my output) so that they can use the data input of Z (processed by AI). AI is thus the enabling technology and gives designers an entirely new field of action and market opportunities.
AI is thus the enabling technology and gives designers an entirely new field of action and market opportunities
A good AI example can be found today in cancer imaging/screening. The AI recognizes by comparing with thousands of learned reference data, if and where in a picture cancer cells. The oncologist conclusively validates this and draw conclusions for the therapy. The value lies in the significantly increased diagnostic reliability.
And what can the designer contribute to that? US based startup Zensors has developed an artificial intelligent software that allows to easily evaluate temporal events, objects and more within a visual action range of video data (say a live video camera). For example, it is quite banal to count the use of serving trays in a restaurant and then automatically deliver new ones. Or gather contextual information to streamline the flow of visitors in it to fill another box office. And this in a quite simple manner with a graphical interface and without Python knowledge.
Every day, more and more weak artificial intelligent software modules are coming onto the market. Like today with Photoshop, Illustrator or Sketch, we will be able to use these modules of data to develop super exciting and above all human meaningful digital product and service compositions. Since we are just at the beginning of the digital transformation, the imagination has no limits.
Maybe then the noise from the flood of data becomes harmonic information.
Originally published by PAGE online in German.