The Variable A in A.I.
There is no doubt the prospect of true A.I. is exciting. It has been projected that the immediately addressable A.I. market is expected to reach $70 billion by 2020, and it is pretty clear most A.I. implementations will have a transformative effect on every facet of society and business globally and being only 1 year away you can understand why it is a hot topic.
So much so that the White House just launched ai.gov to outline A.I. initiatives.
Intelligent Connectivity was the headline at MWC and almost every 3rd keynote and stand at MWC was orientated around some kind of A.I. — Microsoft introduced their A.I. cameras under the Azure Kinect set, BMW demonstrated gaze, point and contextual voice recognition underpinned by A.I., Huawei represented a ‘full stack’ A.I. solution which they anticipate digital transformation in manufacturing, construction, traffic control, retail and smart cities. There were also a number of startups demonstrating their versions of their intelligent automaton services to take R.P.A from macro on steroids to business-ready digital work companions.
Merging A.I. with the physical world was CloudMinds which utilised a cloud-based brain and arms and legs (wheels) of a robot.
Then we had the more fun (or consumer-friendly) A.I. expeditions including more plugins and interactions for Alexa and Google assistant and even an A.I. enabled toothbrush from Oral-B which not only tells me how bad I am at brushing my teeth but will build a health profile on me which can be shared with my dentist, doctor or other physician… perhaps even an A.I. physician eventually.
What all of this actually means, is that a lot of companies all over the world are inspired by A.I. and other data-driven initiatives. These companies are focusing on squeezing as much value as they can from the huge amount of data generated on a daily basis. And for that, they are actually using machine learning and A.I. is actually the output.
Which brings us to the A in A.I.
When looking under the hood of A.I. at MWC this year 3 distinct types of A.I. were being discussed, demoed and promised…none of which would be considered true ‘Artificial Intelligence.
A.I. (Actionable Intelligence)
There were a large number of software developers and algorithm architects offering platforms to collect, normalise and analyse realtime data sets and accessible data lakes to contextually provide the right information and the right time to the right people for human review, decision and action.
Some of the examples on display include actionable security intelligence whereby actionable and comprehensive insight is provided to reduce risk quicker than a team of humans could. In other areas, it was used to provide advantage to businesses by providing realtime insights into competitors and market behaviours, almost like a predictive analytics scenario.
Overall, the interest in Actionable Intelligence was quite high not only because of its ease of use and integration, but also it keeps humans in complete control. The comfort level of attendees around this method was quite high.
A.I. (Automated Intelligence)
The implementation of modern Automated Intelligence at MWC was mainly focused around evolving Robotic Process Automation beyond its current state by adding analytics and minor automated actions.
Many of these examples were about alleviating the mundane, repetitive administrative tasks weighing down humans. This was from simple items such as automatically scanning, processing and actioning invoices (or orders) through to more exciting prospects including mapping IoT and open data in realtime across farms and deploying the appropriate measures for optimal crop yield (based on learning algorithms) which could include irrigation, pesticides or even opening the sheep pen to let the sheep into vineyard to eat the grass around the vines…keeping weeds and growth at bay and letting the sheep naturally fertilise the grapes!
Other implementations were around the automotive and transport areas, letting vehicles self-drive, providing a human was still in the driver’s seat (and awake).
Listening to other attendees quiz the Automated Intelligence providers, it was interesting to see how willing many individuals (and businesses) are to let the software automatically control minor elements of their business, lives and environment on the provision they can regain control at any time, see why certain decisions were made and further train the algorithms with their own datasets.
A.I. (Augmented Intelligence)
The prospect of Augmented Intelligence is by far one of the most exciting and promising near term A.I. initiatives. Augmented Intelligence is designed to complement human intelligence and is by far the closest A.I. we have to the idea of Artificial Intelligence.
Fortunately this does not require surgery or cranial implants.
Humans are facing a big challenge, one that they themselves have created. Thanks to advances in the fields of cloud computing and mobility, we are generating and storing huge amounts of data.
In these data stores lie great opportunities to reduce congestion in cities, identify signs of cancer at earlier stages, help out students who are lagging behind in their courses, discover and prevent cyberattacks before they deal their damage, and much more. But the problem is looking through this data and finding those secrets is beyond human capacity.
This is exactly where augmented intelligence can help humans as augmented intelligence is particularly good at analysing huge amounts of data and finding patterns and correlations that would either go unnoticed to human analysts, or would take a very long time, and presenting it exactly when needed.
On show at MWC were demonstrations that included an AI algorithm which can analyse a patient’s symptoms and vital signs, compare it with the history of the patient, that of her family and those of the millions of other patients it has in store, and help her doctor by giving suggestions of what the causes might be. And all of that can be done in a matter of seconds or less, all whilst the Doctor is engaging with the patient. Likewise, AI algorithms can examine radiology images hundreds of times faster than humans, and they can help human experts in helping out more patients.
Another quite intriguing demonstration was set around a field worker with a set of hololensesq glasses with embedded Augmented Intelligence. In the simulation the field worker was working on a faulty piece of machinery and as the worker viewed the faulty machinery the Augmented Intelligence was ‘seeing’ the same piece of machinery, analysing it and cross referencing private and public data sources including previously recorded faults and their resolutions. Within moments the Augmented Intelligence was showing the human field worker what tools they require, what parts they need and how to start dismantle and repair the machinery along with a probability score that this cause of action would result in success.
Whilst it was a canned demo, it was quite exciting to watch and imagine the wider implications.
There were also AI assistants for retail sales environments whereby a human sales associate would work hand in digital hand with an A.I. assistant to augment the customer service and sales process. The human would engage in a ‘warmer’ way with the customer, read body language and be more social whilst the A.I. would recommend fashion styles based on conversation, observation, spending habits (if known) and what is in stock and automatically ‘scan’ the items through the POS.
Essentially giving every sales associate a personal assistant so they engage with the customer rather than the systems.
The Digital Colleague
This idea of a Digital Colleague who focuses on data and process driven tasks whilst the human is focused on tasks that humans are more efficient at (for the moment) is a driving element of how businesses will start to adapt and change to digital data driven transformation.
In fact the consensus amongst all of the Augmented Intelligence offerings on display at MWC is that AI is not about replacing human intelligence (or humans), but it’s rather about enhancing it by enabling us humans to make use of the data generated by and around us.
This approach and perspective is quite in tune with what society actually wants, or is at least leaning towards when it comes to A.I. InMoment recently released a study which articulate that nearly 50% of Australians think better service from staff is the most important thing brands can do to enhance customer experiences highlighting the need for humans to be ever-present in a digital world, no matter how advanced it is.
And this is it seems is where the near term future is heading, humans augmented by intelligent assistants, it is quite conceivable that within the next 24months there will be a huge spike in employment, just not in the traditional way of employment.
It is not only possible but highly probable that businesses will start employing, for lack of a better word, digital colleagues to augment their staff to increase efficiency, reduce costs and increase profits. And this goes beyond a digital P.A. This is entire teams of A.I. employees who report to a human manager, each A.I. employee filling a new role rather than replacing a human.
It is even possible that the value of the human employee is not only their skills and experience but also the ‘team’ of A.I. digital colleagues they have curated and trained.
But perhaps that is dreaming a bit too far ahead…. what is for certain is that humans and businesses will be able to take advantage of some sort of A.I. and it should not be a question of if, or when but how many and which type.
In our environment at Delv, we use multiple types of A.I.s for different purposes and built on different technologies. We are currently deploying and training digital colleagues to all of our staff to not only make them more efficient but to also enable them for new ways of working. And I see this as one of many steps to true A.I which is more general intelligence rather than speciality intelligence.
As we continue to explore the many permeations of A.I., I implore you to do the same within your organisation. If not into deployment or test at least into a discussion amongst your teams and business units to understand the implications and opportunities of employee enhancement, not a replacement.