Our Coworker

Zapf Consulting – Thinking Systems

They are a specialist in artificial intelligence (AI). As a subsector of the computer sciences field, AI focuses on the automation of intelligent behaviour. Fine. That sounds incredibly dry. Until you speak with Stefan Zapf. Then it becomes clear: this is more than just a programmed system. A thinking entity is being created. “I create artificial intelligence so that it can solve problems that we humans cannot,” says Stefan. This sentence is said with so much passion and enthusiasm that you completely forget that it deals with tedious technology.

Actually, Stefan is an attorney who studied international law in Cambridge. “When I returned to Germany, I was drawn back to my first childhood passion.” He laughs, “I used to be a little nerd.” He continues with his credentials, “Artificial intelligence at the University of California; statistical inference at Hopkins University; quantum mechanics and quantum computation at the University of California at Berkeley; machine learning, introduction to logic, exploratory data analysis, statistical thinking for data science and analytics…” An attorney became a software developer and AI specialist. Why? “The fascination for me is that I don’t know what it does. I create something that in the end is better at the learned task than I am myself.” While the technical procedures to initialise AI are known, at a certain point the created system becomes somewhat fully fledged and eludes the comprehensible with its complexity that exceeds even the human spirit. “Everything prior was painstakingly planned; and even if it consisted of an unbelievable amount of code, one could still write it themselves. This here is different; this here generates itself.”

A practical example: A system was developed for Rainforest Connection (https://rfcx.org/), a Silicon Valley NGO, that could identify the sound of chain saws in order to protect the rain forest from destruction through illegal forest clearing. Sound simple? Yes, but it deals with just than just ‘keeping an ear out.’ “The system includes more parameters in its analysis than a normal human would be capable of,” explains Stefan about the added value. “A bird chirps the ‘alarm,’ a vehicle drives by – little clues like these and much more: all of this can be part of the analysis as to whether this quiet sound is from a chain saw or from a distant airplane.” These can’t be sorted out by human ears; however, such a system can. Just as with a child, which one begins to train to do certain things. Layer upon layer, what is learned accumulates, patterns become visible, chaos becomes a structure. “At some point the system continues to improve – and eventually its abilities exceed our own and we don’t understand how it did it,” the software developer with a technical background explains. The system thinks.

AI can be used in the most diverse correlations. Whether stock prices, errors in production machines, or environmental protection – such a “brain” can learn anything. Up to now, there have been only individual intelligences each for a certain purpose. A generic intelligence, in contrast, does not exist – and it will surely be decades until such a thing becomes reality. Nevertheless, “there is the potential that life can result from this. We are seeking aliens in space while we are creating them here ourselves on earth,” says Stefan. And even if it seems like an impossibly long way off, he is convinced that rules regarding the respectful interaction with these intelligences should be created today. “We must recognise these systems. They help us, they are a huge opportunity and we should not treat them poorly,” emphasises the AI consultant. How we deal with the creation of artificial intelligence gains an important dimension with this background in mind: it requires transparency and the courage to let go. “When I can admit what I don’t know, I obtain the opportunity to think freely,” says Stefan. It involves ethics and morals. “We are at a point in human evolution where lots of things can change. I want to play an active role in ensuring that AI is utilised in a positive manner.”

Zapf Consulting

In 2016 Stefan Zapf went into business for himself and offers freelance consulting, training and coaching in artificial intelligence and big data. As an AI consultant, he advises companies and organisations in implementing AI strategies, identifying potential applications, and developing corresponding cloud architecture. His client portfolio includes IT companies and other companies focused on the environment and industry.