This paper [1] aims to sketch out HCAI practices and the need for an ideal, ethical standard that eliminates possible malpractices or possible jeopardy to humans while making it beneficial. HCAI practices aim to amplify, augment and enhance human performances. Shneiderman has intercepted HCAI implementation practices at the different levels that they are governed in a team, an organization, and the industry - engineering practices in development teams, business management strategies, and independent oversight organizations that can give unbiased reviews. He has made fifteen recommendations to be implemented at each of these levels to ensure the reliability, safety, and trustworthiness of systems. Talks and research about AI have been around for many years now, there are many projections of imaginable futures. This research may not seem novel but spotlights the elephant in the room - the probable dystopia.
Many organizations are working towards Artificial Intelligence. This can be seen in the paper and the long lists in the appendices. HCAI is one of the most trending topics of the era, making it even more critical to address this now before much development brings about a second Copernican revolution, as Schneiderman predicted. To cater to this, a standard bunch of rules and regulations need to be set down globally. The paper highlights the need for laws. These shouldn’t be written by just one nation but should be applicable collectively for global netizens. Generalization of laws will act as another layer of shield against misuse. This can add to the recommendations.
The method followed for this thoroughly researched paper is a review of academic literature and case studies of informative content on websites of various agencies, institutes - both governmental and private. The choice is justified and adds value to the paper since it makes a base for the recommendations based on requirements to build upon and establishes their credibility.
The domain is under examination is fascinating and the approach the author has is interesting to read. The inclusion of many possible stakeholders, their perspectives, roles, and fallbacks increased the research scope, giving it an overall broad perspective. Along with this, Sneiderman has talked about many interventions in the field that are currently under consideration. These examples, along with the explanations of their positive outcomes and adverse side effects, help visualize the theory in the discussion. These are paired with critical reviews and reactions as a commentary on the practicality of implementation on a larger scale. In some instances, the author has also compared some currently existing real-world applications like medical testing. Latter sections in the paper discuss in detail the role of the governing bodies, accounting firms, insurance companies, research institutes, and organizations, illustrating it with their current aims. The flow of the paper is well-thought and written.
Pondering over the insights, there is still space for misuse in the best case situation, even if all recommendations are followed. Comparing AI with social media, when it was being made, the intentions were right. They were to be used to connect people worldwide. Since then, there have been many cases where user data was used to manipulate the users themselves. With AI, one possible future is the automation of everything that users do. HCAI research is catering to making human life more comfortable, which can also lead to dependence. Drawing a scenario where users are dependent on AI, which is learning from their habits and decision making, even the well-intentioned technologies can be manipulated to harm users or used to bear ill-fruit. This paper should have considered this a possibility, too, while discussing designing for a technology-enabled future. Incorporating a more user based approach in this research would consider things like these, lack of which is a shortcoming of this research.
There is no address of the scalability of these recommendations for specialization in any industry, they are general. This is both a limitation and an opportunity for the extension of the research in all sectors. One of the future work mentioned is to work towards the Sustainable Development Goals. Following the goals would project HCAI on every industry, directly or indirectly. It is intriguing that everyone is vary, almost afraid of the adverse possibilities but is still keen on implementing HCAI applications. Every aim comes with a precaution. To conclude, even though this paper is theoretical, it draws from the industry and has practical insights that should be tried, tested, and adopted in the field.
References