Challenges in using bot computation to ethics of the Web – By Aditya Abeysinghe
Respecting users
A web ethic that is commonly maintained is respecting users. Respecting each user by others during meetings, informal communication, and formal communication is often maintained to ensure that each party is not disrespected and insulted. By respecting others, qualities of communication and human qualities are enhanced which may cause better information to be exchanged.
However, bots (chatbots and other bot types) used for websites and web-based apps have little to no intelligence on how to respect users as they just transfer data between two ends and use trained models based on machine learning and other analysis techniques. Some bots may improve how they exchange data during communication by learning and updating their inner model used. However, emotional qualities of bots are few compared to that of humans. This has been one disadvantage of bots as most people seldom opt to use them to request or exchange information due to lack of emotional qualities.
Ensuring security
Security is another issue when using bots. As bots use input provided by a user, they could easily be used as entries to attacks. Text can be input to bots to create an attack as there is less validation for texts which are used for attacks. Also, bots lack ability to reason and avoid abnormal input such that attacks could be avoided. Humans could identify when input they receive is abnormal and input only data that is not harmful to systems.
Privacy
Users often need to ensure privacy of data captured when they use chatbots. They would consider that data including identification numbers, and Internet Protocol (IP) address are not captured while using chatbots. However, as processing methods used cannot be viewed users often have less trust on providing such data to bots. Also, there are fewer filters on how bots output data from internal sources. They might expose sensitive data when processing and responding to a user.
Lack of knowledge on expressions
Chatbots often use models where user requests are converted to a system understandable type. They often use data from a source and process results and respond using these models. The issue with this method is that when users request information that is unavailable in these sources chatbots cannot provide responses back to users. Also, as bots have less ability to express their feelings on such issues, users often get low user experience with the website or the brand.
Image courtesy: https://www.investopedia.com/