Artificial Intelligence- thoughts on a moving target
The challenges of adopting Artificial Intelligence (AI) are posited as many, but in fact there is fundamentally one. The challenge is the preposterous delay in waiting. What creates this wait? Nothing really, just the febrile nature of our imaginations, sustained by countless science fiction renderings of how artificial intelligence and our species’ future plays out. The pendulum swings wildly from from blind optimism or abject terror, depending on your tastes. The point is that these extreme portrayals of humanity, engaged in an endless war with technology, has led us to this now widely accredited inflection point in the development of our species. Never have we needed to look as deeply at the effects and penetration of technology in our lives. Is artificial intelligence a step into something too personal for many to comprehend, the genesis of intelligence itself? What is our basic understanding of the term? It reflects on us how much time we need to stop and consider what we mean by intelligence. Industry has data attached, and intelligence beyond algorithms is not discussed. As a consequence, we have been locked into thinking that artificial intelligence is fundamentally about productivity and not for the way that it changes us as we come to consider it in our lives.
Let’s start at the beginning. Machine learning represents the root from which AI has branched. At basic level, data structures are fed exemplars from which input sources (from streams of data through to computer vision trained to recognise faces) can work with algorithmic structures that mimic the neurological modus operandi of the human brain. From this, outputs can be generated – speech synthesis that can respond to linguistic stimulation (chat bots), or human-like pattern spotting of key inconsistencies of legal contracts, in another example. What unifies these is that emergent properties of information become linked into what appears to us as knowledge, a key pillar of what we denote as intelligence.
There are several challenges ahead for business. One is how to bring this intelligence to bear on the kinds of business challenges faced when much of the processing power exists in high power cloud based computing facilities and depend on good data connection to be applied. Another is the very specificity of business challenges, which can be a network of factors, from market development through to supply chain integrity. How does a business owner know what to plug into any form of AI?
Take an industry such as logistics. The challenges might be conceptualised thus:
Gaining access to existing datasets and AI outputs, for instance the usage of traffic flow information that could help schedule the outgoing courier slots to best work around a series of urban routes between source and destination
Working to create seamless offline/online experience for drivers, so that any frustrations in using an AI to help moderate your work rate does not meet a wall of resistance. Many other mobile based experiences engage clever interaction paradigms to achieve this, and AI will have to do the same
AI as it informs automated processes, this pertains to how AI might, for instance, moderate production in a robotics based manufacturing business. If given scope, AI might influence product development and production cycles and this has far reaching consequences for growth based businesses
Internationalisation, still not the strongest suite for AI, and in businesses with transnational operations, there needs to be a candid understanding of what decisions might need to be referred up to executive decision makers, based on adapting to cultural nuance
AI not appearing to be intelligent, being able to dial up or dial down levels of intelligence speaks to the comments at the start of this article about what might actually be meant by intelligence. If a fully formed answer is expected, anything less would perhaps lack fulfilment and perceived usefulness. The connection between impact on the bottom line and outputs from a form of AI (given the investment in infrastructure or subscription services needed) is likely for shareholders