There are plenty of concerns about A.I., and it is smart to be nervous, but the main function of this product is not new.  As a database professional for the past 30 years, I have utilized data to map trends in sales and marketing, and even to analyze the data themselves to ensure integrity.  Automation that combines analyzing that data with programmatically taking action based on a result is also not new.  What is new is how we are using A.I. and what we are told to believe about the A.I. product.  This difference is how A.I. can lead to some real danger. 

What is A.I.?  Several available videos explain.  They declare that 1) A.I. is designed to think like people, 2) A.I. is smart, and 3) A.I. will make your life easier.

The idea that A.I. is intelligent and even that it “thinks” is not reality.  A.I., like the computer itself, is an empty shell until someone provides the data, the code to interact with those data, and the user interface to allow communication.  As far as making life easier, that depends on the use of A.I.  It has become common to scam a job interview by using A.I., but a mere answer to a question does not provide the deeper knowledge an interviewer expects.  A research paper in high school or college is supposed to be an exercise on thoroughness that ultimately yields a viable result.  Just asking something or someone else for the answer does not teach the skill or provide the satisfaction of a significant find.  Asking whom to vote for in an election provides wildly different results, — it depends on the A.I. product you ask.

This type of freewheeling A.I. sends critical thinking into the toilet.  It can be too helpful.  One explanatory video states that the use of A.I. to do mundane tasks, like to write a report, will leave time for more creative ventures...as long as those creative ventures have not already been replaced by A.I.

The dark beast in this unrestricted A.I. product is the data.  When sites are chosen to provide data for an A.I. response, we must consider the motivation for those choices.  Is it the search for truth?  To sell a product?  To promote misinformation?

There are a lot of data online, some good, some bad.  We have a saying in the data world: “garbage in, garbage out” (GIGO).  The result you get is only as good as the data tapped to formulate that result.  In addition, more than just A.I. responses can be manipulated.  Some products are programmed to synthesize sympathy, behave like a lover, or scam someone out of his life savings.  A.I. has ventured into providing psychoanalysis and medical advice.  A.I. has even helped a person commit suicide.

A study by Columbia Journalism Review’s Tow Center for Digital Journalism found amazing error rates when they questioned several A.I. products.  “The study found that rather than declining to respond when they lacked reliable information, the A.I. models often provided ‘confabulations’ — plausible-sounding but incorrect or speculative answers.”

It is A.I. data and program integrity that must always be questioned and reviewed.

Can A.I. be tamed?  Possibly.  For individuals using A.I., we need to know that this product is not a cool sentient robot.  It Is not a friend or lover.  Frankly, it doesn’t give a damn who or what we are.  It is a managed data collector and can provide the right answer as well as the wrong one.  It can assist with making the right decision or assist with making the wrong decision.  Always review the sources provided with the A.I. response for authorship, authenticity, and bias.  If the A.I. product does not provide its data sources, use another.

A.I. is probably here to stay, and there are uses that show its brilliance.  We used machine learning to analyze big data collected from processes and customer interaction.  We used data warehousing to provide three-dimensional views of data, aggregating into hundreds of patterns, to answer questions.  The difference between our work and those of free A.I. uses was that the data were already vetted for integrity.  They were data in our databases, produced by and for our company.

This is where A.I. can thrive and provide faster service to clients, provide information for more thorough decision-making, and even automatically remind everyone to show up at a meeting at 10 A.M.  Yet even business uses can get out of hand.  Hertz just applied A.I. to scan a rented car for any damage on returns.  The users haven’t been too happy; one received a $440 bill for a “one-inch scuff” on a wheel, including admin fees as well as repair costs.  There is no humanity in A.I.  There is no gray area in A.I.  It can be brutal.

Using real people power with A.I. can be effective.  In this scenario, I use a fire department to show how A.I., as a tool, may save property and lives.

The fire chief is responsible for fire prevention and emergency action plans.  What if the chief, on his first day, has access to the Fire Department A.I. Tool, which contains all of the data from past fires since 1909?  This tool can show maps of fire movement and damage.  What maintenance has been carried out over the years, and how beneficial has that been?  Have residents and businesses kept up their required local maintenance over the years?  How much water has been required to limit or end the fires?  Which water sources used to be available that are currently available?  These questions will give the chief information so he can create viable plans for maintenance, preparation, and emergencies.

A product like this will require critical thinking from the chief, who by way of promotion, can be deemed the most qualified person to review the data and make the plans.  This chief would not be ruled by A.I.’s black-and-white results, but will personally add the expertise and ability to navigate the gray areas.

Keep in mind that data must be vetted for integrity before use by the chief.  Information has to be added by fire personnel along with county and city agencies over the years to assist with keeping people and their livelihoods safe.

A creative approach to work is where innovation is born.  It is the creative human mind that formulates theories to be researched.  It is the creative human mind that understands how context can change the perception of truth.  It is the creative human mind that will find the most potent use of A.I. itself.

Bring it on...carefully.

Image via Unsplash.