bend, blend, break, build…

September 30, 2019

Chris Duffey is author of Superhuman Innovation: Transforming Businesses with Artificial Intelligence, and Head of Artificial Intelligence Innovation and Strategy at Adobe. He spearheads Adobe’s Creative Cloud strategic development innovation partnerships across the creative enterprise space.

Amidst the multiple debates about its efficacy and perceived superiority vis-a-vis human intelligence, AI seems to assume a ‘non-human’ dimension. What is perhaps being lost upon many is: the focus has to be on how to combine AI with human ingenuity so that the technology can be used creatively for the betterment of the world.

There is a saying attributed to Albert Einstein that the true sign of intelligence is not knowledge but imagination; and this is a foundational theme when we talk about a human-centric approach to artificial intelligence (AI). In the case of AI, the true measure of intelligence is intelligent imagination. What we mean by that is imagination can now be more informed and more tangibly enlightened through the unlocking of data. As such, artificial intelligence is the platform of our time, the medium of the moment because businesses can now triple down on the AI value exchange to enhance innovation.

When it comes to AI, we do not only have to understand the technical aspects, but also need to focus on how and where it can be used and leveraged to solve business problems. A relevant metaphor is that of the maestro or the composer. They understand the capabilities of the instruments within their orchestra, but they do not necessarily need to be able to play each of them or even know exactly how the instruments were created or constructed. Their role is to bring all those instruments together to create a masterpiece. That is essentially what we are talking about with artificial intelligence on a business strategy level. Customers and consumers are not interested in the technicalities of AI. They want to know what it can achieve and how AI technology can serve people and businesses.

An additional vital fact to remember is: none of the AI technologies are intended to replace human beings. Instead, humans and artificial intelligence will work together to create a brave new world. A world where people are freed to use their natural creative abilities and their amplified intelligence without concern for the drudgery of mundane, repetitive and, quite frankly, boring tasks. This is the true mission of artificial intelligence. And because of this, the possibilities for the betterment of humanity are endless.

AI has been called a ‘suitcase word’, in that it contains many smaller concepts that can be unpacked and analyzed. This essentially means that the phrase ‘artificial intelligence’ has such broad meaning and application that there is not just one way to understand it. Today, AI can be thought of as ‘the study and design of intelligent agents’, which are systems able to perceive their environment and act based on what is happening. The subject of AI overlaps with computer science, data mining, facial recognition, robotics, and others such as the study of the human mind.

Today, AI can be thought of as ‘the study and design of intelligent agents’, which are systems able to perceive their environment and act based on what is happening.

To fully understand artificial intelligence, we must reference human intelligence. A lot of people think intelligence is measured by a thing or a test, but actually human intelligence can be categorized into nine types, according to Howard Gardner’s Multiple Intelligences: New Horizons In Theory. These are intrapersonal, spatial, naturalist, musical, logical-mathematical, existential, interpersonal, bodily-kinesthetic, and linguistic. Of course, everyone has all these components to a greater or lesser degree. Some are strong musically while being weaker spatially, while others are strong in the existential area but are introverted and thus have less in the interpersonal area. However, people are holistic in nature – they do not normally fit into just one category. Herein lies the complexity of artificial intelligence, if the purpose of which is to try to re-engineer aspects of the human mind.

With that general understanding of human intelligence in mind, there are essentially three forms of AI: narrow, general, and superintelligent. When we talk about modern AI, we are talking about narrow AI, which means artificial intelligence that is designed to perform specific tasks. Internet search is an example of an AI-powered discovery task, and it has become ubiquitous in its use throughout the population. AI chatbots and Q&A algorithms are another example, they can answer customer questions, assist with customer service and help customer representatives with suggestions about what would be most valuable to the customer. The label ‘narrow AI’, however, does not really do justice to what AI can do. Contrary to the implications of the term, the capabilities are vast.

Narrow AI can also be thought of by role instead of technology. That [of] an assistant, peer, or manager. The first layer is that of the AI assistant, an example of this are semi-autonomous cars that include a human driver for more involved driving situations. Mobileye is an advanced driver assistance solution that helps drivers avoid accidents. This company was founded in Israel and the product is presently installed in many public buses, helping human drivers to avoid accidents. Systems such as the Audi Traffic Jam Pilot and the BMW Traffic Jam Assistant help by taking over steering, braking, and accelerating during heavy traffic. Because they are intended only to assist, they require you keep your hands on the wheel even when they are maneuvering. Another example of an AI assistant is the KLM Customer Service Rep Tool. This AI-powered deep learning application helps customer service agents deal with ‘the overwhelming volume of messages coming at them via social media and other channels’.

An example of AI operating as a peer can be seen in sport play predictions. Two students at North Carolina State, William Burton and Michael Dickey, built a system to predict whether an NFL team would pass or run. In a 2014 game featuring the Cowboys versus the Jaguars, the model was right 91.6 percent of the time. AI can also be used as a manager. In 2015, commuters in America spent over 8 billion hours stuck in traffic. Stephan Smith, a robotics professor at Carnegie Mellon University, equipped traffic lights in Pittsburgh to use AI to react to traffic patterns as they changed. By doing this, travel time was reduced by 25 percent. Airlines generated $168.2 billion in revenue in 2016, and traffic is projected to double in two decades. Because of this, airlines are looking into how AI can help them keep up with the demand and improve their service. A core benefit of this symbiotic relationship is that humans and AI learn together and from each other, in addition to each improving over time.

By putting the human first and creating new experiences, companies have to leverage distributed cognitions which are frameworks to allow and celebrate the free flow of ideas which can up-level humanity to be even more creative. The technical term for this is computational creativity. In the creative professional world for instance, there is a spectrum of creativity: production, execution, ideation, and inspiration.

By putting the human first and creating new experiences, companies have to leverage distributed cognitions, which are frameworks to allow and celebrate the free flow of ideas that can up-level humanity to be even more creative. The technical term for this is computational creativity.

Presently, computational creativity is good at specific tasks in production and execution. However, there is debate if AI can ever truly ideate and be inspired. The point being that AI is here to amplify human creativity while helping on specific tasks. There is a lot of interest and excitement happening with GANs. GAN stands for generative adversarial networks and are ‘deep neural net architectures comprised of two nets, pitting one against the other (thus the ‘adversarial’). These were introduced by Ian Goodfellow at the University of Montreal in 2014. GANs are important because ‘they can learn to mimic any distribution of data,’ including images, music, speech, prose, and anything with unique attributes and characteristics. Examples include everything from creating anime characters to posing three-dimensional images to creating new backgrounds on videos and movies.

Simply, two different neural networks combine to work together to create different content. Let us say you have a base photograph and a reference of a Van Gogh painting: the GAN can essentially render your photo in the style of Van Gogh. That is a simple example, and conceptually the technique can be applied to just about anything. But we must remember these are not replacements for people; they rather are tools to amplify and augment our natural abilities. Technology exists to help people and further the goals of humanity, business and so on. Without humanity in the picture, the systems have no purpose.

With collective intelligence, creativity can be used for business invention. Human ingenuity plus AI will bring about invention across energy, fashion, finance, pharmaceuticals, recruiting, retail, advertising, art, automotive, aviation, banking, energy, security, and sports. We must first learn the rules of human-AI alignment, so we can creatively bend, blend, break, and build upon them. Ultimately, we have to keep in mind that the end goal is not the technology, it is about what humans can do with it; in this case, what people can do with it creatively for the betterment of the world.

With collective intelligence, creativity can be used for business invention. Human ingenuity plus AI will bring about invention across energy, fashion, finance, pharmaceuticals, recruiting, retail, advertising, art, automotive, aviation, banking, energy, security, and sports.