AI is changing the web forever

Before I move on to explain this rather big statement – and maybe convince you of my approach – we’ll need to take a closer look at the state of AI.

AI was coined in 1956 by Dartmouth Assistant Professor John McCarthy:

“AI is a general term that refers to hardware or software that exhibits behaviour which appears intelligent.”

AI has the potential to tackle profoundly difficult problems, and the solutions to those problems can be applied to sectors important to human wellbeing.

Netflix uses AI driven algorithms to keep its viewers engaged and reduce cancellations. Tesla combines their autonomous cars with rich databases regarding roads, traffic lights, weather conditions and other relevant rich databases. As a technique and concept, in Tesla’s instance, these developments may lead to increased road safety and decrease traffic congestion. But it’s the combination that is impactful. The real impact lies in the ways in which we will utilize all this data.

The point I’m trying to make is that AI clearly is in its infancy. Google’s image captioning tool is another example of this. It’s trained to recognize objects in a photo with up to 93,9 percent accuracy. It exemplifies a brilliant use of data, but the limitation lies in the intelligence of the model. The model is trained by using human captions, and will thus only be as ‘smart’ as it is trained by human input. 

Both these examples reveal that in order to push AI further, we need to invest less time in programming and more on training the applications. Businesses are competing for the most powerful (smart) databases and therefore we need to find smart ways to acquire rich databases, and think of how machine learning can utilize this data in ways that the human brain might not be able to. We are thinking too narrowly when discussing AI. If we keep defining AI as a tool or a service, we are by default limiting its true potentials.

Instead, we need to look at it as a new paradigm. Think about it this way: we are currently designing for hardware (mobile / desktop) and software (apps, web, intranet, and so on), because these are the current ‘homes’ for this type of content. But looking at it from a historical perspective, we are moving from the industrial revolution (the replacement of physical power to machine power), to the second machine age (replacement of machine power to brain power). And within this second machine age, we are shifting from the mobile-first age to an AI-first paradigm. If we look at AI this way, its true potential can be unleashed.

Deeming AI as a new paradigm forces us to think bigger. It will urge us to refocus and think more about establishing the parameters of this new paradigm, such as building qualitative databases, unifying design patterns for AI driven tools and create more out-of-the-box AI solutions to work with. Finally, considering AI in this way pushes brands to understand that investing in AI is a necessity. Because if it really is the AI-first age, and brands are not investing, it will massively reduce their competitive advantage in the long term. 

Brands like the idea of AI. But what we see is that they end up with ‘easy solutions’, like chatbots, instead. Let me explain. Brands invest in chatbots because it’s easy to convince yourself that investing in chatbots equals investing in AI. The messaging interface is perhaps the easiest way to experiment with AI. In my opinion, brands don’t fall in love with chatbots but the potential of intelligence the chatbots represent.

Because yes, chatbots are currently not intelligent. They are a conversational interface trend, but are a long way from human-level conversational abilities. In all fairness, chatbots tend to disappoint customers – for whom they are developed – because of this. In my opinion, we need to stop developing chatbots that are supposed to act as humans. Take them for what they are. Present them as what they are. So, if we agree that AI is a new paradigm, it becomes clear that chatbots are not smart.

Chatbots could be smart, and in the future, I suspect we shall see it develop as a much more integrated service. This circles back to my argument in the opening analogy of this piece; in today’s internet-driven world, the sky can often be seen as the World Wide Web. But Artificial Intelligence (AI) is about to change that forever. We need to consider that AI interfaces aren’t necessarily based on screens - nor the ‘limiting’ experience of the web the screen or mobile-first paradigm offers. Websites, for example, are ‘dumb’ because user-flows and outcomes are predefined. AI offers a different spectrum completely. Adding AI to websites can help predict user needs and create opportunities to give new meaning to an existing platform.

To further explain my point, I need to talk a bit about Virtual Reality (VR). VR is changing the way we’re designing experiences because there are no ‘physical’ framework limitations (browser-screens). The same is true for AI. When combining VR and AI, you clearly see how current platforms can get new meaning. Brand communities, for example, are currently located on existing social channels (Facebook, Instagram, and so forth) or owned channels (forums, corporate websites etc). These are mainly driven around member-interaction and content (2D). In the near future, brand communities will be translated to VR experiences where brands can actively participate and be present in the community via smart chatbots or other forms of AI (3D). What’s different is that in addition to the content and conversation, the ‘physical’ sphere and personality of the brand will have an impact on the experience of the visitor.

It will be interesting to see where these communities will be located in our digital landscape. Is the web ready for smart 3D developments? YouTube (Google) is focusing on 4k 360 videos whilst Facebook launched a chatbot marketplace for messenger. The relevant question to ask here is who has the most powerful (and therefore smart) database? With a smart future approaching fast, it is no longer about reach but more and more about relevance.