**3. Artificial intelligence in user Interface design**

Design focus with Artificial Intelligence paradigms in User Interface is referred to as Intelligent User Interface (IUI) design. AI is revolutionizing industries and changing the status quo, and the design of UI is not left out in its disruption. AI is being used to design tools for providing details to UI designers to aid their designs. AI tools for User Interface (UI) with focus on User Experience (UX) like Uizard and Airbnb's Design AI can turn design sketches into product prototypes [19]. The interfaces must be intelligent enough for adequate detection of cyber threats. An AI-based User Interface threat model will assure reduction in cyber risks. A mainstream example of IUI is the GMail Smart Compose that offers predictive text to complete sentences thus aiding in writing emails faster [20]. An Intelligent Virtual Assistants (IVA) or AI assistant is another example of AI in UI design generally called IUI. Some notable IVA are Amazon Alexa, Microsoft Cortana and ChatBots. Generally, AI can be applied to UI designs on various interfacing channels between human and machines especially when modeling cyber security threats. Channels like search engines (Yahoo search, Google search, Bing, etc) and content recommendation systems (Netflix, Youtube, Spotify, etc) are also significant examples of IUI. Data privacy and considerate integrations, among others are challenges encountered with AI in UI designs, and they must be factored in during the design process for cyber security treat modeling [21]. IUI is aimed at incorporating intelligent automated capabilities and Artificial Intelligence (AI) into Human Computer Interaction. They are human machine interfaces that aims at improving effectiveness and efficiency of Human Computer Interaction (HCI) through the use of reasoning, user models, domains and media such as GUI, VUI and Gesture-Based User Interfaces [22]. IUI sits between AI and HCI. **Figure 1** depicts the relationship between AI and HCI with consideration to IUI.

**Figure 1.** *Relationship between AI, HCI and IUI [23, 24].*

HCI provides design techniques for user interfaces that are efficient and the AI components are used to embed intelligence into those interfaces. Basically, IUI is used to depict interfaces that generates some sort of output or exhibits some behavior in which the user interacting with the system considers the system intelligent [25]. A typical example is when a user clicks a wrong button and the interface is able to guide the user on the right button to click. Also, assistance to the user must be readily available through IUI.

Due to diverse models from differing fields that greatly influence user interfaces, the development of IUI is no different, as there are disciplines such as AI which contributes the intelligence in simulation approach to improve responsiveness; software engineering which allows for formal language definition, unified modeling approach, development life cycle; and the Human computer interaction that evaluates the user experience and avails techniques that can be used to create usable user interfaces [26]. In addition to these, there are other fields that still significantly contribute to the development of IUI which is depicted in **Figure 2**.

IUI tends to enhance the interaction between computer and human by bringing about novel approaches of communication and adapting the interface to the user using techniques of AI [27]. The concepts of IUI tend to the intersection of the area of AI and HCI in the field of IUI [28]. This is with a view to automate users' task while engaging in the threat assessment on cyber security [7]. Therefore, the application of intelligence in UI design will in no small measure bringing about better monitoring and controlling of activities that enhance better cyber security measures and threats modeling applications.

The advent of computers which primarily uses a mouse and keyboard as a medium of HCI has brought about a new dimension in the design of UI. This gives rise to the creative and innovative way of UI design for cyber security tools smartly through the use of AI. A new generation of UI which is called IUI keeps rising day by day and improving UX in HCI. IUI attempts the integration of intelligence in HCI that enables the automated capability of UI with the view to improve usability that enables high performance of HCI in cyber security software. It employs design and

**Figure 2.**

*Research fields of the intelligent user Interface [27].*

## *Application of Artificial Intelligence in User Interfaces Design for Cyber Security Threat Modeling DOI: http://dx.doi.org/10.5772/intechopen.96534*

implementations of AI components that perceive, learn, interpret, reason, and decide with a view to enhancing the capabilities of cyber security experts effectively [24] in decision-making. A different approach and tools can be employed in the implementation of IUI design. The choice of tools and approach is however depending on the goal and the area of application in the cyber security modeling tool. The strong effect of AI on UI/UX is that it helps fix challenges and avails insights and discoveries, however, humans have to recognize the challenges first. AI is therefore less likely to be pre-programmed and more likely to be built with technology so that it can fix the challenges by sustained learning and based on this learning, decisions can be more customized and user-centric. Data can be analyzed and reported in real time with AI.

Applications of machine learning (ML) that communicate directly with daily users are now progressively diverse and ubiquitous. ML is designed to allow a computer to learn about the past or the present and to forecast or predict the knowledge for unknown events in future [29]. The technology's sophistication and prevalence facilitates the belief that AI is the new UX; that is, AI would be the most effective way to boost the user experience [30]. This is because AI will permit UX designers to personalize contents through the application of machine learning algorithms to build intelligent models to mitigate cyber threats.

Typically, the development and design of UI occurs over the following phases as shown in **Figure 3**,


However, using AI, a handwritten design can be translated to a working UI as shown in **Figure 4**. This code generation usually go through a training phase and a sampling phase. The training phase is where the model is trained to identify images and state the relationships between them using algorithms like the Convolutional Neural Network (CNN) and Long Short-Term Memory (LSTM) and the sampling phase is more like the prediction phase where samples of UI to be translated will be supplied to the model so as to evaluate and fine tune it's performance.

**Figure 3.** *UI development phases [31].*

**Figure 4.**

*Translating handwritten design into working UI [31].*

The benefits of using AI in the design of UI include but are not limited to:


AI systems can establish a deep bond with humans by creating relationships. Relationships foster where trust grows and this relationship could either be human to human or human to machine. One of the major reasons for the wide adoption of technology (especially AI) is the growing trusts in these systems. This reliance is strengthened by a repeated show of ability in achieving purpose, speed, accuracy, usability, security and privacy amidst other things [32]. By mining enormous data, through the use of deep learning, the trust between AI and humans can be strengthened by the continual generation of useful insights that fits into diverse purposes

*Application of Artificial Intelligence in User Interfaces Design for Cyber Security Threat Modeling DOI: http://dx.doi.org/10.5772/intechopen.96534*

and ensure goals are achieved with speed, accuracy and security thereby increasing the usability of such systems. Considering the example of phishing site, the user's trust in the system is bound to increase because the system has protected the user from the phishing attack. Therefore, the user could rely on the AI system to detect phishing sites and help mitigate cyber threats. The more we interact with AI systems, the deeper the connection. For every interaction, data is generated and added to an individual's digital footprint on a system which AI feeds to improve its intellect with humans. AI technologies are also evolving rapidly and currently can analyze human emotions, sparking wild discussions into its unforeseeable future [33].

#### **3.1 Empirical studies on intelligent user interfaces**

The complexity and the number of cyber attacks is increasing [34] on a daily basis which poses a lot of threat to both public and private technological assets. Cyber security employ threat modeling techniques in mitigating attacks against technological asset by hackers. Threat modeling allows proactiveness and provides insight to risk evaluation and prioritize mitigation [35]. Formal methods have been found to be profound for performance evaluation compared to traditional methods [36]. As the cyber threat is becoming prominent, application of Artificial Intelligence (AI) in cyber security is another advancement in the technology employed by the experts in the field [34]. The manifestation of AI - enabled UI are Jarvis, Amazon Alexa, Neflix, IBM Watson, Nest Thermostat, Spotify and iRobot Roomba [37]. While advancement in technology to enhance security measure against cyber threat is expedient and important, User Experience (EU) also need to be given a high priority in the User Interface (UI) design to enable high usability and hassle-free workflow. As better security of system should not be tantamount to worse UX, the important factors for enhanced UX and UI in cyber security are balancing the security of system/software with UX, designing of UI based on human perception and minimizing the complexity of software integration into the existing network infrastructure [38]. A poorly designed UI reduces UX which leads to a user performing desired actions with difficulty [39]. An application of AI plays a key role in the IUI to mitigate problems that may arise from human interactions with machine. IUI is regarded as a subset of Human-Computer Interaction (HCI) research with a goal to use smart and current technology to improve HCI.

Applications of AI on UI can drastically improve the interaction between humans and computers. This makes it possible for computers to understand more human communication channels like body gestures, hand gestures, sounds, eye movements, lip-sync and other body motions [40]. This advanced communication with humans via AI has led to innovative solutions addressing human-computer communication barriers. An example is Conversational AI, which is an advanced platform for the widely used online service helpdesk. Conversational AI can analyze an individual's emotion and manage frustrations on the system by routing to different channels for better customer service [41]. With IUI, adaptive, personalized and responsive services can be provided to ensure the specific need of the user is met even when they are yet to realize it. With big data, data mining and deep learning algorithms, technologies that can drive these personalized services could be developed to work in synergy with intelligent environments to mitigate environmental challenges that may arise as a result of human computer interaction.

As the important features of IUI is to enhance the HCI, the following are current techniques used in IUI [27]:

1.**Getting of input from user intelligently:** This involve innovative way of getting user's input through several techniques such as recognition of face and expression, processing of natural language, recognition and tracking of gesture, and tracking of gaze.


The application of intelligence in UI design is to improving users' experience that enables efficiency, effectiveness and user satisfaction using different approaches. This is achieved through the representation of reasoning or acting in accordance to a set of models such as user, dialog, domain, tasks, or speech. As shown in **Figure 5**, different models from different disciplines constitute the development of IUI. AI contributes the simulation of intelligent techniques to enhance the communication, software engineering- enabled notations, unified processes and formal languages while HCI deals on consideration pertaining to users. These combinations of models allow the creation of techniques that allow usable user interface.

IUI applies AI techniques to different input and output with a view to improving UX intelligently. This is achieved through reasoning, representation of knowledge, machine learning, adaptation, and adaptivity as exemplified in various applications such as email filter system, dialog system, email response system and so on [43]. As threat modeling provides answer to questions, "where, what, and how" [3], so also the metrics derive from these questions serve as input and output in AI to improve UX and integrate intelligence in UI designs.

There are two different methods of a user interface that can effectively be applied to intelligence user interfaces in cyber security threat modeling namely direct manipulation and indirect manipulation [7] as depicted in **Table 2**. Also, IUI have been anticipated as a way to overcome a number of the issues that direct manipulation interfaces cannot deal with and further highlighted three principles that can be applicable to threat modeling in cyber security, namely (i) control transparency and predictability (ii) Privacy and trust (iii) treating systems as fellow beings [47].

**Figure 5.** *Disciplines involved In the development of IUI [42].*

*Application of Artificial Intelligence in User Interfaces Design for Cyber Security Threat Modeling DOI: http://dx.doi.org/10.5772/intechopen.96534*


#### **Table 2.**

*Comparative analysis of IUI.*

Ehlert [27] also highlighted a number of issues that recent direct manipulations interfaces will not be able to handle for instance creation of tailored systems, filtering problems, provision of assistance on the use of complex and new programs, taking up the responsibilities from the users with other forms of interaction. The characteristics for developing intelligence in user interfaces designs that enhance the development of a threat model in cyber security are:

