Author: Dr D Dhanuraj

Trustworthy AI (Artificial Intelligence): A Tool Rooted in an Objective and Balanced Approach

To create trustworthy AI (Artificial Intelligence), we need to rethink how autonomous systems interact with society and with other AIs, which differs from how humans and computers typically work together.

Transparency and responsibility must be a top priority in governance frameworks and design architectures. This is especially true for engineers who work in non-traditional areas where incentives often conflict with moral obligations. These systems, which may act in ways that are not typical for humans (for example, risk-seeking algorithms that prioritise accuracy over caution), need to be carefully examined for biases that are built in, whether they are inherent in the design, the performance measures, or the way they balance different values.

Trustworthiness depends on the reliability of the technology and the system’s ability to handle moral uncertainty effectively. This is where philosophical problems arise: Should AI incorporate human biases from the past to reflect cultural contexts, or should it transcend them through neutral logic? Dual-use tensions make things even more complicated, so governance needs to focus on data stewardship instead of strict data hoarding.

It is strange that making information more accessible to everyone has made expertise less valuable. National policies should encourage competition through open data ecosystems rather than hindering technological progress. To balance different priorities, such as safety versus innovation and problem-solving versus mitigation, it is essential to employ an interdisciplinary approach that incorporates philosophers, social scientists, and coders to understand the reinforcements, historical biases, and positioning, thereby ensuring neutrality and objectivity.

To create trustworthy AI, we need to rethink how we evaluate values in system design and determine how to balance the need for independent efficiency with human-centred ethics. This is a balance that has been refined over hundreds of years in technical, legal, and cultural contexts.

The paper was originally published by the Friedrich Naumann Foundation.

Dr D Dhanuraj is the Founder-Chairman at the Centre for Public Policy Research (CPPR), Kochi, India.

Views expressed by the authors are personal and need not reflect or represent the views of the Centre for Public Policy Research.

Also, read our paper on “Generative AI and its Influence on India’s 2024 Elections“