My Stories

DownloadTowards Singularity: Implications to Intelligent UI with Explainable AI approach to HCI

Intelligent User Interfaces (IUIs) have emerged as a field which involves designing and implementing an AI that overly covers human skills and capabilities, so that human per- formance with an application excels. AI is on a verge of recursively improving and gaining meta-intelligence. Machine learning, Robotics, Automation Technology, Artificial Intelli- gence, and Internet of Things are the disciplines that will play a chief role in unifying human knowledge and in developing an Intelligent Interface. Gender differences and dis- abilities also generate an important need for developing a better UI. The recent trends in technological advancement have seen improvement from traditional mouse, keyboard devices (WIMP Interfaces) to various Tangible User Interfaces (TUIs) which has helped in getting closer with the users. IUIs include email filters, email response systems, spoken dialogue systems, and head- worn User Interfaces. Intelligent processing is very difficult to predict and it involves limiting the responsibility of users in their thoughts and judgment as it can outperform traditional human techniques. As of now, humans are intelligent narrow-predictable systems. Re- searchers are working towards understanding the spontaneous unpredictability of humans through IUIs. Currently, Intelligent User Interfaces are available more in the User Interface of the system compared to the intelligence available in the backend of the system. The communication between the users and the system can be enhanced by implementing multimodal interactions, natural language processing, graphical interfaces and conversational mediators. A need to replace traditional, screen-based HCI techniques and interfaces with new collaborative learning interface emerges.

Explainable Artificial Intelligence (XAI) proposes to make a shift towards more transparent AI. It aims to create a suite of techniques that produce more explainable models whilst maintaining high performance levels. [29] Along with developing Intelligent UIs, explainable capabilities should also be incorporated. AI systems are getting more complex and XAIs can result in making AI more understandable to humans. Along with transparency, trust concerns also play a vital role when decision making capabilities are handled by the system and XAI acts as a formal technique to collectivize the social, ethical and legal pressure calls. Undesired results are often yield by AI based systems. XAI acts as a measure of justification, control and improvement of current AI based systems. For users to understand the model, the interaction between the humans and machines should be effective. Proportionality thesis, which states that increases in intelligence lead to proportionate increases in the capacity to design future generations of AIs, is true. [26] With advancement in training algorithms a possibility of technological singularity can be agreed. AI systems might recursively improve itself in order to reach dominance. Quantum computing has drastically reduced the limitations of computational abilities and also influences hypercommunication. But there is always optimist and pessimistic approaches towards this concept. Predictions of the end of the human job because of replacement by robots and AI are lacking in sufficient analysis and evidence that cover the technical, social and economic effects.[27] However, designing futuristic UIs will surely need improved and self learning AI capable of generating adaptive interfaces on its own. Cyborgs can also help in shortening the gap of connection between users and the system thereby creating reliable and appropriate datasets for generating intelligent interfaces. The paper has been structured across sections of impact by singularity. Firstly, the recent technological limitations are described. Next, the social implications of intelligent user interface is reviewed, while presenting the contributions from Artificial Intelligence. Finally, a general overview of proposed design is described in depth. The paper ends envisioning the economical and medical implications towards singularity.

Image placeholder

Sushrut Ghimire

Researcher, Active Learner and an engineering student.