**5. Artificial intelligence, Big Data and dynamic policy: Policies inspired by natural systems**

Before we proceed to the model, a minor detour is inevitable as a foundation for the proposal. This detour briefly outlines the concept of artificial intelligence (AI). A growing discussion in the world of computer science is around the types of computer intelligence. Terms that include, cognitive computing, machine learning and deep learning, are the focus of this chapter around artificial intelligence (AI). Often, there is an interchangeable use of these terms in daily language. However, the terms differ although refer to related things.

Cognitive computing refers to the sensory subdivision of machine intelligence that is used. Sensors and algorithms are used to enable computer to 'see', 'hear' and 'feel' [35]. Through image sensors computers can see, microphones facilitate their hearing ability and the text-to-speech and speech-to-text technologies permit computers to interconnect with humans using natural (human) language through programmes such as Alexa, Siri, Cortana and Google Assistant.

Simplified, machine learning is knowledge computers gain from old data and historical trends identifying patterns that humans cannot identify [36]. This form of machine intelligence uses colossal amounts of data to generate patterns recognised by computers and thereafter used to differentiate objects from each other, for example, distinguishing between male and female humans or cats and dogs including the different breeds of these animals. The deep learning branch of machine intelligence involves using neural networks that mimic the physiology and function of the human brain [36, 37]. The networks include several layers of neurons that permit computers to learn from historical data and thereafter apply in a way similar to how a human brain thinks [36]. This is the most advanced form of machine learning which is increasingly becoming the favoured approach in training computers.

AI refers to machines acting in ways that seem intelligent [35]. This is through enabling decision-making capabilities to computers. The intention of AI is for

computers to make decisions that address specific problems just as humans do every day [35]. Computers use recommendation engines for this purpose, whereas narrow AI is a machine-based system designed to address a specific problem such predicting election results [15, 35, 37]. AI applications work in several branches that include machine learning, natural language processing and robotics. Our proposal is to extend the application of AI into the policymaking space to inform climate change mitigation and adaptation based on models adapted from natural systems through the process of biomimicry. This is a proposal for a dynamic policymaking model. The quintessential proposition of the model is the sequence of quick interventions that incorporate rapid and automated feedback loops that reinforce learning in the policy cycle so that the policy remains in a state of dynamic self-improvement as shown in **Figure 1** (the Mjimba-Sibanda dynamic policy model).

The first stage of the model entails an analysis of natural biological systems using the biomimicry lens in order to understand the strategies that nature has evolved over aeons. The appropriate and relevant strategies are turned into design principles that are no longer limited to the biological context. These general design principles are used to inspire the development of a base policy. To update the policy continually, whenever defined factors change, the abstracted principles are modelled to produce algorithms that mimic the behaviour of the natural biological system, including in terms of having specific variables for the algorithm. The algorithm's input is Big Data from both the public domain and relevant databases for public policy. Artificial intelligence leverages the sensing of relevant data input, computing the historical and live data in order to adjust the policy and to provide feedback to improve the modelling of the algorithm. Most attractive is that computing the Big Data can enable the prediction of future scenarios. In principle, this means avail an opportunity to anticipate future challenges and adjust the policy to avoid or adapt to those changes before they even manifest. Adjusting to avoid such changes is the essence of climate change mitigation, and adjusting to manage the impacts of the changes relates to climate change adaptation- anticipatory adaptation. The desired policy adjustment can either be automatic, which is desirable in some cases, or be an outcome of debates that are characteristic of the democratic policymaking process.

**Figure 1.** *The Mjimba-Sibanda dynamic policy model. Source: Authors.*

*Biomimicry, Big Data and Artificial Intelligence for a Dynamic Climate Change Management… DOI: http://dx.doi.org/10.5772/intechopen.84406*
