Five Ways AI is Driving Digital Design and Convergence


These are questions that design-focused teams need to ask to stay competitive with upcoming technology products. One of my team members loves to quote Johnny Mnemonic cyberpunk sci-fi author William Gibson: “The future is already here, it’s just not distributed very evenly.” It’s true. Artificial intelligence (AI) is already changing the landscape of digital design and user interface paradigms.

“What’s the difference between Machine Learning (ML) and AI? If it’s written in Python, it’s probably ML. If it’s written in PowerPoint, it’s probably AI.” It’s a popular joke among programmers, poking fun at how loosely AI is used, often by startup founders just to get the attention of venture capitalists.

Anything ML or AI is growing fast, so let’s take a look at their impact. Can they help people create better digital designs and user experiences? What trends can we observe today? What ideas can we use for our future products?

1. Better design for accessibility

I’ll start with accessibility because AI can make a real difference in the user experience for people with disabilities. A few years ago, Facebook published some amazing technology as part of its alt text project. It used object segmentation, facial recognition and voice generation to help blind people “see” photos by touch. By moving their finger over a photo, a blind person would get a voice explanation of the object like, “Johnny is smiling. It’s dark and he’s wearing sunglasses. This experience wasn’t available before the era of the AI.

2. Unbiased Design

Can AI help us create less biased design? Reduced human involvement in decision-making processes reduces bias. We can teach AI to treat people fairly and determine if design is biased. Big companies spend a lot to maintain a clean reputation (with many eyes on their successes and failures) to do just that. AI could be an important step to help this process.

However, we have to be very careful about the data the models use to learn. If we feed a machine learning model with biased datasets, the bias will be amplified. Forbes contributor Mark Sears described COMPAS as a prime example of bias in this area. article. With COMPAS, racial bias was inadvertently included in the algorithm to predict the likelihood that criminals would commit repeat offences.

3. Trademark protection, plagiarism and fake ID photos

According to Global Trademark Counterfeiting Report 2018, online counterfeiting was estimated to be responsible for losses exceeding $300 billion in 2017. Current ML algorithms are very good at identifying patterns, like analyzing MRI images for cancerous tumors or identifying leopards snows in photo databases. Likewise, AI can be useful in uncovering illegal use of trademarks, logos, and digital art. See what TrademarkVision is doing with brand recognition. AI can also identify if a digital photo has been faked or photoshopped by analyzing color data, unnatural contrasts or silhouette changes in an image.

4. From special effects to AI-generated art and animations

Over the past few years, we’ve seen several AI-first products add artistic special effects to our photos and videos. Prisma apppopular in 2017, transformed images in the style of Van Gogh or Monet. Can AI generate proprietary art based on its own “sense” of creativity? I believe it will happen sooner than we think, and it will surprise us more than Banky’s self-shredding art.

Photo editing is another area where AI can save designers a lot of time by removing noise from photos, enhancing low-resolution images, and even restoring missing objects. What about video? AI can help create super-resolution 8K content, helping to generate things like smooth slow motion and realistic light rendering, even enabling AI-generated animations and people. Have you seen the Chinese news anchor 100% AI powered? This is just the beginning!

5. Predictive and personalized user interface

The best UI is not a UI at all. If the software we use can analyze and understand our needs on the fly, then why would we still need the user interface? Today, software uses Zero UI. It predicts our next wish based on incoming data and our lifestyle habits, like when Apple Maps gives us a deadline to get home, even when the app isn’t open. The more data we provide, the more predictive the UI can become.

And our world is full of data, including various Internet of Things (IoT) sensors, wearable devices, computers, phones, home assistants, cars and things that haven’t been invented yet. Netflix is ​​a well-known example of a custom UI where it selects a still image or video slice from a movie, along with a description based on your previous history. Netflix believes this significantly improves conversion rates.


According to Forrester (via Forbes), it is estimated that investments in cognitive computing systems will amount to $1.2 trillion by 2020. Even so, we are still in the early stages of AI development. But the faster AI and technology evolve, the faster they will both accelerate. We are already wondering when will AI achieve its own sense of creativity, create its own designs and code its own software.

It’s not far. It’s already clear that designers have to think completely differently just to keep their edge. This includes thinking about ways to remove bias from data, keeping tabs on how data is being used, and how this big data can be put to practical use. That’s terabytes and petabytes of data in the world right now – so much data that it’s far beyond our abilities as humans to interpret it all on our own. Considering how a The AI ​​managed to cheat in carrying out its tasks, we may be forced to find answers to a much wider range of questions faster than we would like.


Comments are closed.