The human challenge for Artificial Intelligence

While AI is being touted to be the all-inclusive solution and productivity enabler tool for the next generation of industrial reforms it is most likely be limited by the ability of human mind to comprehend and analyse information in real time basis or the inherent unavailability of all the information at the same time to be able to take cognitively coherent decisions. And doing that consistently is another challenge!

• Different cogs in the chain are always running at different speeds and humans need to feed enough data to the machines in the right formats initially and then continue to provide that feed as the world around changes at an accelerated pace. Machines will also keep learning from the feed and develop their own optimized solutions but conflicting information from internal systems and external world will be a challenge to resolve in the long term.
• Environmental/External impact is unforeseen and cannot be built into probable use cases; while AI with time will be able to predict some, “unforeseen” is unpredictable by definition and supply chain is the most apt place where Murphy’s law applies in unison with execution.
• Building right data ahead of time (such as cost) is factually not possible leading to sub-optimal decision making by any tool or human. With previous similar examples what humans develop as instinct cannot be coded tangibly into AI limiting it’s ability.

Availability of information and flow of information both are important but more important is trend of that information over time; in a rapidly changing environment data sets are morphing which leads to loss in mapping of information and hence trends over time; limiting the ability of machines to continue learning over time and drive decisions. Humans on the other hand can quickly develop mental mapped perspectives (often laden with biases); even on broken data sets but the limitation of human mind applies!

Humans need a new mindset to work along with AI; that skill set is currently very scarce! The only solution in the short to mid term is that AI being used as a secondary tool for checking on data and cross-referencing decisions. However, when the tables turn is anybody’s guess!