I asked myself this question about a year ago after I started Data Literacy Norway - a Meetup group with a data scientist friend.
When data literacy is defined as the ability to read, write and communicate with data in context, it is legitimate to ask whether you can read artificial intelligence, whether it makes sense to learn to write it and how you communicate with it. The answer is not necessarily obvious from the point of view of the non-expert.
Yet the answer is yes. Artificial intelligence is part of data literacy. It relies on data to deliver analytical services. Therefore, it is necessary to know how to read and interpret the results it provides. Understanding its foundations is also a good idea: how it is developed and how it works. However, for the user of AI services, it will be difficult to check the quality of the data, the relevance of the algorithms or to understand the possible biases. From a data literacy perspective, the challenge is to know how to ask the right questions while acquiring the necessary skills to understand the answers.
Before going further, let's have a look at the types of services an AI can deliver. It can be diagnostic, such as biometric identification or detection of a patient disease based on a series of indicators related to medical history, lifestyle or environmental factors. It can also be used for predictive analysis, such as forecasting the weather, anticipating a risk of disease or a natural disaster for an insurer, or anticipating the needs of its customers for the digital marketer. Finally, AI can help automate tasks or prescribe recommendations to increase productivity. In logistics, for example, it is an AI that will organise a deliveryman's round according to the time available and the number of packages to be delivered. AI is also used to build chatbots or the virtual assistants in our homes such as Siri, Alexa or Google Assistant.
We can clearly see, through these few examples, that AI has infiltrated all areas. We are increasingly interacting with algorithms and for this reason we need to have an understanding of what data means and how it is processed.
Here are some questions to address:
About AI and working processes - no matter where you work:
How much of my work is managed by algorithms or could/should be?
How is the competence related to data collection and processing shared across the organisation?
Within the organisation, how does the transfer of competence regarding the development of algorithms or their use occur between experts and non-experts?
What is the level of collaboration between data experts and the rest of the organisation when implementing artificial intelligence?
Have working processes been adapted to the use of AI?
About the origin of the data that feeds the AI and the purpose of its use:
Which data is collected by the tools I use, e.g. my voice assistant, my car, my bank card, the surveillance camera, the sensors in the fabric of my dress etc.?
How will this data be used? By whom?
How does the algorithm interpret the data?
Does the data provided to the algorithm come from reliable and consistent sources?
How long will the data be stored? Is this duration justified?
Just as knowing how to read a book or a newspaper article allows you to question the form and content, knowing how to read data allows you to question the form and purpose of an algorithm. You may not always get the answer, but knowing how to ask the right questions is a guarantee of greater transparency and ethics. It is also an opportunity for experts to discover errors or biases and to correct them. The more data literate we are within organisations and as citizens, the more accurate and efficient artificial intelligence will be. As the proverb tells, forewarned is forearmed.
