Tuesday, January 16, 2018

AI Automation Versus AI Assistance

We keep hearing about the momentum of AI and the various facets of AI. AI is hot, but what does it mean for organizations and people? Will AI automate our jobs away? Will AI be the friendly assistant that multiplies our given skill sets and competencies? Will AI be a foe once it realizes how weak humans can be? Will countries leverage AI for defense or offense purposes? These are all questions that keep people engaged or up all night. This post is just one debate that will occur in organizations.

The Case for AI Automation:

Organizations are almost always looking for ways to cut costs while making sure their resources are well managed. AI is great when embedded in robots to handle dangerous tasks, so that is almost a "no brainer" for automation. AI is also good embedded in special purpose robots that save labor over the long haul. We have all seen the videos of robots in warehouses or on production lines where speedy precision is the goal.

These a pretty obvious and proven use cases, but there is far more to tap. There are more complex tasks that are ripe to eliminate. Quality assurance focused jobs that leverage sensory focused AI are ideal for automation. There a goodly number of white collar jobs that are about mining and analyzing data and information. Where speed of pattern sensing is key, AI automation makes great sense. Over time AI will be involved with selecting the analysis techniques that will be applied. More and more of today's human labor focused jobs will be automated except where soft skills are needed. Chatbots are even eating into some of these tasks.

The Case for AI Assistance:

While it seems clear that low skill level jobs or ones that require significant speed to action completion are ripe for automation, it is also clear that AI has a great role in assisting humans in completing their tasks as well. There are a goodly number of low level assistants that can handle simple natural language tasks including fast dictation, information look up, calendar scheduling, and travel advice . These are practical, but low level task support.

Additionally, AI can be used to devour information in a speedy and predictive fashion while incrementally learning faster than most humans. This allows AI to advise humans as they sense emerging patterns, sort out multiple contexts and understand an array of complexities. There are software bots that can handle specialized informational and analytical tasks that can push labor to lower levels of skill. Imagine putting experts next to your people to make them smarter. Examples of human collaboration includes medical advice to underwriters, medical treatment plans for doctors and diagnostics for mobile service professionals.

Net; Net: 

There will be more emerging examples of both automation and assistance leveraging AI. I don't think it is an "either or" situation,  but a combintorial "both" situation. Automation is an econmic necessity, but so is the augmentation of humans to handle more complex tasks and shorten the expreince curve. The debate will shift over time and the exmaples will get more complex but hamans and AI will continue to coperate. Everyones skill levels will have to rise for people over the long haul.

Additional Reading:

Myths of AI

Top Seven Uses of AI