Potential users should carefully research the use of AI and ML to properly set their expectations. As with any technology, unless and until you recognize the limitations as well as the pros and cons of it, your choice of direction might not produce the result you desire.

Recent studies of visual facial recognition AI use note several facial attributes and coloration tend to reduce the accuracy of the correct identifications. This type of bias extends to accents found audio transcriptions, which we can observe if you talk to Alexa or Google Assistant or Siri and have a local drawl or speech pattern that the algorithms cannot interpret. Learning systems can adjust and compensate for some characteristics, but accuracy suffers.

I would want to know what the business owners application will be. If using AI/ML for digesting large amounts of data to use as the basis for market research and decision making, then you would want to be able to train on the parameters that are most meaningful for your situation. If tagging images or performing speech to text or text to speech applications you want the seed sample size to be sufficient to be able to drive longer term learning of the engine.

The list of considerations always comes back to the application, and the expectations as to whether the technology is ready for prime time production use, and of course the cost considerations to produce a successful application that returns value to the investing business.

Quick checklist for business owners to consider when applying artificial intelligence to their business -

1) Clarity of purpose on why AI -- Business Intelligence / Analytics / Collaboration / Complex tasks / Training / Monitoring / Any other?

2) Outcome expectations after AI implementation --- Cost Savings / Reducing human workforce / Productivity Gains / Make life easier / Better Safety / Drive predictions / Any other ?

3) Ensure to avoid algorithmic bias in the chosen AI implementation

4) Check if the AI implementation can be achieved within existing compliance framework in place for the business

5) Understand implementation cost. Ask the right questions before decision to move forward

First they have to consider for what they need it, they need a strong business model. Furthermore, they have to be aware that AI requires adequate Human Intelligence to work together. This not only means IT-experts, but a general workforce, which is open to work together with the intelligent algorithms.

Personally? I find the whole AI debate a little disconcerting! Call me Old Fashioned, but just the concept of replying on computers to determine my clients' needs is slightly worrying. However. Equally, and nonetheless. I am always quite interested in technology generally, and so I'd like to think anything machine led has a place in developing my business. Perhaps I'd give it very basic, routine tasks to begin with, to see how it goes. Does this add value? Regards, Jason