The title of my presentation can only hint to the ambiguity. An answer requires three clarifications: what is the objective, what is the setting, and when is this happening.
Today, organizations that wish to implement AI face five challenges: strategic, structure, functional, cultural and ethical.
However, depending on the AI maturity this will most likely change. While, at present, prediction seems to be at the heart of most AIs, in the future, when judgement, autonomous decision making, or even self-conscious and emotional AIs become available, different challenges will evolve as well.
After a short introduction on how we got to where we are, we will quickly touch on what is currently possible and what we can expect mid-term, before we dive into the 5 challenges involved, when driving AI adoption in an organization.
Asking, what an AI would do, given a certain objective, in a particular setting and at a given time in the future, we learn about Intelligence-driven Process Automation (IPA), an implementation method that’s just as agile as long-term focused, allowing to harness the benefits of AI implementation while still experimenting.
Finally, we contemplate a world where AIs develop themselves and ask again: what would an AI do?
One of the main questions people ask about AI is about substitution: will come the time AI will undertake tasks humans do?
It's very important to understand the difference between applied and generalized AI.
Today and in the next 15-20 years we are talking about applied AI, means only special kind of tasks will be fulfilled by AI, but also when in maybe 20 years generalized AI will come (a machine will "think" like a human) there will be enough creative and "other" tasks for humans so that AI will not replace, but could extremely transform humanity for the better.