AI – Artificial Intelligence has many controversial definitions. This can make it tricky for organisations to understand how AI fits into their processes.

In a broad sense Artificial Intelligence is research into how machines can carry out human actions autonomously.

Is this possible? I don’t entirely believe so but that is why AI is research.

However, when organisations think of AI, they are thinking of specific problems which they would like to be more efficient on.

Problems such as the mundane aspects of their operations where improved efficiency, improved accuracy , speed of information search and so on can be achieved.

In the instances listed above, “Artificial Intelligence can be thought of as automating processes to achieve improved efficiencies”.

Or rather an improved description would be ; Organisations “use” Artificial Intelligence to automate processes and achieve improved efficiencies.

Are we Using or Doing AI?

The word “Do” is a verb which means “to perform, take part in, or achieve something” – Cambridge Dictionary

The word “Use” as a verb means “to put something such as a tool, skill, or building to a particular purpose” – Cambridge Dictionary

What is Using AI? Using AI is the implementation of Artificial Intelligence research outcomes . For AI to be used, it needs to have been researched, developed and tested. The concept that various components of AI can be orchestrated to produce outcomes/results needs to be developed. If we want our AI code to recognise a human male or female, we need to train the model with a vast amount of information. Only then will it be successful. The probable viability of the model is also assessed. After this, AI is adopted and used by all.

Most use cases we have been approached with require the use of AI. This is indeed the right approach as not all organisations can do(research and build) AI. It is not realistic nor is it feasible for every organisation to “Do” AI.

What is Doing AI? Doing AI involves the research into how machines can be programmed to perform tasks quickly and efficiently, generally for the benefit of humans. Doing AI requires extensive computing knowledge, data processing and computer programming knowledge. Various programming languages are used to develop AI models such as python, C#.

Contrary to the implicit assumption that AI replicates human intelligence, one AI doesn’t do all different things at once.

A human individual can generally, listen (Speech recognition AI), speak (Speech synthesis),perform calculations(Generative AI), reason and adapt themselves. AI solutions require specialised and synchronised AI models/components which communicate and collaborate efficiently to provide the magical outputs of efficiency.

Doing AI requires training the machines to recognise data using metadata. To allow the machine to provide outputs which are intelligent, the AI input requires a large amount of data for it to be trained. These large amounts of data requires expansive datacentres . Not all organisations however can afford nor justify the cost of such data centres. It is not necessary and especially when there is no business case to do so. The research which has been done by Computing giants who specialise in research is sufficient for organisations to Adopt AI and use AI for their needs. It reduces footprint and is beneficial for the planet.

There may be a time, as the adoption of AI progresses, where many organisations choose to develop their own AI. However, significant research is still needed. This research will enable this realisation, especially concerning data centers.

Similar to the advent of computers. When the computer was first introduced, the machines were huge. They could not fit into homes. These machines were required to perform simple operations and calculations. Fast forward to now, we have computers performing complex processes in our homes and hands as phones and so on.

Image Credit : Getty/Historical

For many organisations to “Do” AI themselves, data storage will need to evolve. Computers have achieved this evolution. We are not quite there yet with data storage.

Should your Organisation Do or Use AI?

It depends…..

Most organisations will find that they are “Using” AI. The Use of AI should be adopted when there are appropriate business use cases.

However, when adopting the use of AI, this should be done in concert with Governance and Change Management. This will ensure that the organisation is able to see positive returns from its use of AI.

References: Doing AI – Richard Heimann, Metadata – Jefferey Pomerantz