• Artificial Intelligence is like Lego; to build something nice, you need to combine the right pieces in the right way.

    Most of us have played with Lego when we were small. I did at least and I absolutely loved it. I can remember the days when my friends and I were playing with Lego for hours on end, constantly creating new structures and building complete cities. We loved it and I am sure it stimulated my creativity. Even adults still play with Lego, often in group exercises to get some creativity flowing. Lego is a wonderful product and I thank Ole Kirk Christiansen for inventing it.

    Lego actually comes in thousands of different shapes and 100 different colours and with that, the most amazing structures can be built. Each Lego piece can only be used in a small number of ways (depending on the size of the Lego block). Each piece is compatible with every other piece ever created and every Lego piece created is manufactured to an exact degree of precision, with a fault tolerance of only 10 micrometres. With the right pieces and enough creativity, you can build almost anything you wish.

    Artificial intelligence is no different to Lego; you want to make sure that different algorithms are compatible with each other, you want to make sure that the algorithms are correct and have minimal fault tolerance and when you start to combine different algorithms, you can create an algorithmic business with enormous potential.

    Different algorithms can be used for different applications and each algorithm should really focus on one thing and be extremely good at that. Building a generic algorithm that is good at everything is, not yet, feasible and it is likely to remain like that for at least the next decade. Therefore, if you wish to incorporate artificial intelligence into your organisation, you should start combining different algorithms in different ways to solve a problem. And to manage this AI Lego building process, your organisation requires a Chief AI.

    Why You Need a Chief AI

    As with Lego, if you wish to succeed in building something awesome, it helps to know in advance what you want to build and which pieces you require for that. The same goes for AI; which business problem do you try to solve and what technology pieces do you require for that. As such, the business need should always be the driving force, as otherwise, you end up with either the wrong puzzle pieces or a different solution. In both cases, you have spent a lot of resources without addressing the actual business need.

    Therefore, organisations should prevent that AI transforms from a means to an end, to the goals itself. As Kristian Hammand, chief scientist of Narrative Science and a professor of computer science and journalism at Northwestern University, argues, the Chief AI should not bring “the hammer of AI to the nails of whatever problems are lying around”. Specific problems require a specific solution with specific data requirements and to manage this difficult process, you need a Chief AI.

    Job Description of a Chief AI

    That is why a Chief Artificial Intelligence Officer should have a thorough understanding of the business, combined with a thorough understanding of the technology. The Chief AI should combine an understanding of the technology with an understanding what those technologies can do for the organisations strategically. As such, the Chief AI will need to have several important responsibilities combined into one person, rather than splitting them up across the board. The four key responsibilities of the Chief AI are:

    1. Act as a liaison between developers and strategists

    The Chief AI should be able to act as a liaison between your developers and the strategy team. He/she should be able to foster a team where different team members have a strong fit when engaging with each other, yet who can easily focus on their own respective domains to achieve the business need. Building a team where developers and strategists are aligned is difficult, as usually, the two speak different languages. However, such alignment is key to the success of AI projects.

    2. Attract the right AI talent

    Since the Chief AI has a clear understanding of the business objectives of the organisation as well as the available technology already in-house, the Chief AI should be able to attract the right AI talent. By knowing which puzzle pieces are missing when solving a business need, the Chief AI is in the right position to hire the best talent for the job. Next to attracting the right AI talent, the Chief AI should be able to retain this talent by offering them interesting and challenging AI projects.

    3. Understand your organisation’s data needs

    Without data, there is no artificial intelligence, since especially deep neural networks require lots of data to train the algorithms. However, this data should be unbiased and of high-quality, requiring the Chief AI to understand data governance best-practices and be able to apply these to the internal data processes. With artificial intelligence, garbage in means garbage out, so ensuring the right data governance practices is an important task for the Chief AI.

    4. Be business savvy to understand business needs

    Above all, artificial intelligence serves the business needs and it is not the business serving the AI needs. Therefore, the Chief AI should be able to understand the business needs and be able to translate these to technical requirements and adapt (existing) AI tools to the business needs. It is the Chief AI’s responsibility to always keep the strategic business objectives in mind when integrating artificial intelligence into the business.

    Final Remarks

    Artificial intelligence is like Lego; different algorithms can be combined in different ways to solve different problems. Different algorithms have different strengths and weaknesses and the Chief AI should manage the process of offsetting the weakness of one AI with the strengths of another algorithm and vice versa. So, for an organisation to become an algorithmic business, it is more about the application of artificial intelligence rather than the development of new tools and reinventing the wheel, which is why your organisation needs to hire a Chief AI.

    This article originally appeared on Datafloq.

    Image: patat/Shutterstock.com