Innovation Management Internet of Things

Artificial Intelligence for Individual and Collective Efficiency

16 June 2016


Artificial Intelligence for Individual and Collective Efficiency

A Recent History of Intelligence in a Digital World

Artificial Intelligence is a technology that uses human-like learning to perform tasks. The idea of Artificial Intelligence or AI is nothing new. As a concept it has been in our literature and art for centuries. But these ideas had no foundation other than as philosophies of nature and science fiction. But things have changed.

Artificial Intelligence has two models of operation:

  • Deterministic or ‘rules-based learning’
  • Connectionist or ‘machine learning’

Connectionist paradigms of Artificial Intelligence have a somewhat more flexible approach than their rule-based cousins, but both have applications in modern technologies. Machine learning is based on artificial neural networks, which are a much-simplified version of how our brains work. Rules based systems are based on a set of pre-configured rules that are based on pre-existing knowledge, like an expert system. Machine learning creates neural pathways based on incoming information, which gives them flexibility and scope. Deterministic methods can create highly specified systems that determine outcomes from existing knowledge. New methods, such as fuzzy logic are also being employed – creating a hybrid model known as ‘neuro-fuzzy’, this helps to improve the accuracy and flexibility of Connectionist models.

It hasn’t been until the latter part of the 20th century that our technology has been able to allow us to truly utilize AI. In the 1990’s we saw the coming of age of Artificial Intelligence with major investment in technology and the advent of the Internet; this opened up new business channels and opportunities. We started to see machine learning used in data mining, games and translation. By the late 90s, AI started to mature and ‘intelligent agents’ were being developed by Professor Pattie Maes. These agents were used for many tasks, especially repetitive ones, such as web matchmaking, used for example in online shopping comparison sites.

As time progressed and Internet connectivity matured, we entered a new wave of change and AI came into its own. This new era brought with it opportunities in the form of technology growth and product needs. One of the most well known uses of AI is in the world of translation, an example being Google’s AI based translation engine. Google Translate originally worked using a rule based, Deterministic AI system, which anyone who used it back then will remember resulted in less than perfect translations. Google now uses the ‘Tensorflow’ machine learning algorithm for translation. Google translation still isn’t perfect and does not replace a human translator, but it is much improved through the use of a more sophisticated Connectionist AI mechanism.

And today, AI is having a major renaissance with analysts Gartner predicting that Advanced Machine Learning will be a strategic technology in 2016.

2016: The Year Where Big Data Meets AI

In the world of Artificial Intelligence we are witnessing an alignment of planets where big data and AI meet. Data is being generated at a rate never before known. According to market analysts IDC, the amount of data being generated is doubling in size every two years, and by 2020 we will be generating 44 zettabytes annually. This is due, in part, to the explosion of the Internet of Things (IoT) and the ever-increasing connectivity of everything.

The amounts of data we are generating are not only big and getting bigger, but they also include highly complex data sets, needing deep analysis. The type of information that is being generated is the collective knowledge of the planet and offers a rich seam of insight.

At the same time AI, a technology that utilizes data, is maturing. This combination of mass information, working with a technology that can interpret and use these data is a very powerful mix. This will open up new product areas and deliver exciting business models. Artificial Intelligence is the best technology to take the masses of data generated within a modern connected world and use it to interpret, model and predict. This is a powerful and exciting combination, which can fully exploit the data explosion.

The Ghost in the Machine

The data we generate represents a massively interconnected web of information about who we are, what we do, when we do it and how. It is the collective knowledge of the globe and the best way to exploit that is with help in the form of a machine that learns.

Artificial Intelligence, however, is just that, artificial. As human beings continuously learn throughout their life, so AI based products will still require some human intervention to modify and guide the neural pathways. AI is really an extension of who we are, rather than a replacement, and as such, will become an intrinsic tool within many domain areas, whilst remaining under human control. In this way, many areas of AI are ideal for human-machine collaboration. Auditability of AI based products is key to the continued improvement of such products. Auditing of AI pathways and decision-making will be, in itself, a task suitable for a human-machine interaction.

And what is to come? AI is causing a positive disruption to the way our businesses work; the front office will be transformed by the use of AI. Indeed the front office itself will utilize AI to manage all of the complex operations from start to finish. Productivity will be boosted, but this will all be done as a process, across multiple layers – this process needing to be carefully orchestrated; to create the Information System for the world of tomorrow, the human operator needs to ensure it is organized in successive layers.

AI needs humans: The simple recognition of the appropriate data used for processing by an AI system will require human expertise in the first instance. The choice of which type of AI to choose will be dependent on the task; some tasks can be done using sophisticated Deterministic AI, especially those that require a wizard type interface, with known outcomes, but more complex, predictive modeling needs will require a more powerful and novel AI approach, one which fuses human-AI collaboration.

AI is a disruptive technology: and they worry us, we are always going to feel like they present a risk to a business. But they also offer opportunities for growth. History has shown us that the pace of change in our world is increasing. By October 2014, less than 10 years since the introduction of a modern smartphone, 64% of U.S. adults owned one. The same is expected with AI – what is novel today will become de facto tomorrow.

AI will be a two-way street: Continuous improvement of AI outcomes will be a human led operation, whilst the result of an AI operation will improve business operations, health outcomes and so much more. In the world of AI there is a fundamental parameter: to choose the correct cognitive approaches based on desired results and rules.

Ultimately, the human being is the ghost within the AI machine, challenging us to refine and optimize our own decision-making processes. But the AI-human collaboration is a case of the whole being greater than the sum of the parts.

Investing in the Future: The Future is Now

The global market for AI products is estimated to reach around $23.4 billion by 2025. This is because AI is an exciting space to be in and offers many opportunities. It is a disruptive technology that has the power to not only change technology, but also offer massive improvements in existing technologies. This view is backed by the surge in investments into artificial intelligence products; the CBI reporting a sevenfold increase in VC funded investments in AI startups since 2010.

Artificial Intelligence is a very different flavor of technology. It is not as straight forward as the server-client architecture we are so used to. Even the disruption afforded by the Cloud cannot be compared to the implementation of neural networks. However, the opportunities it affords cannot be missed. AI cuts across all technology areas, its applications being highly diverse. For example:

  • In the medical field, AI can be used to analyze data generated from medical imaging that can then be used to model and predict pathways of disease.
  • In the legal world, contract writing has been massively optimized through the use of AI – a typical contract taking a human half a day and prone to errors, whereas an AI written contract takes 15 minutes.
  • In banking, AI is used to detect fraudulent activity as well as more effective speech recognition.

Sopra Steria: A Keystone in the World of AI

Sopra Steria understand the potential of Artificial Intelligence and its multitude of applications; we know that using AI is more than just picking a product. Sopra Steria’s deep knowledge of the application of AI means that we can identify business processes that can directly benefit from the integration of AI and generate models for that use. This takes the guesswork out of choosing an AI solution. As mentioned earlier, the use of AI is still based on human-computer collaboration and Sopra Steria are able to identify, manage and support the use of AI through our expertize.


Leave a comment

Your email address will not be published. Required fields are marked *