Solving the AI Governance problem: What should institutions look out for?

In DEC Executive Briefing #006, “Solving the AI Governance Problem”, DEC President Danny Bielik interviewed Professor Douglas Arner, Kerry Holdings Professor in Law at the University of Hong Kong. In this article, we share an extract of Professor Arner’s insights on AI governance and best practices in higher education institutions.

‍

The Key Issue: Accountability

Questions about AI use are not limited to just one realm of operation, but stretch across the entire organisation.

How should students be taught how to use AI? 

How can AI be used in human resource processes, keeping in mind concerns about personal data and privacy?

How can academics use AI appropriately in the research process?

How can AI be used in human resource processes, keeping in mind concerns about personal data and privacy?

‍

In the interview Professor Arner identifies accountability as the main concern in AI governance and a guiding answer to the questions above raised about AI use. When making use of AI tools it is important to ensure that there is someone who is ultimately responsible for the results of AI use. The key concern for institutions will be ensuring accountability for AI use throughout the institution on its different levels.

In the field of academic research Arner foresees accountability as less of an issue due to the rigour and structure of research processes and the expectation for research findings to be accounted for, whether aided by AI or not. He contends that the standards for such accountability must continue to be upheld by institutions.

When it comes to AI use in assessments both faculty and students should be ready and able to challenge or justify any results derived through the use and aid of AI tools.Institutions should ensure that students who use AI tools in their assignments and assessments are ready to take ownership of, and responsibility for the results and findings derived from AI.

This leads us to an important consideration - who should be accountable for AI use in the organisation? Furthermore, how should accountability be assigned?

Arner expects accountability for AI use to follow existing standards and processes in institutions. In many cases, institutions have a designated committee or person in charge of AI use and governance who are then accountable for the organisation’s policies on the use of AI.

“The AI Idiot”‍

Arner also warns against the potential pitfalls of unchecked over-reliance on AI tools. The best outcome from AI comes from the combination of AI tools and human expertise. “The AI Idiot” is a person who defaults to using AI tools in place of developing their own expertise, causing us to lose out on the greatest benefits of AI.

The next generation, current students and fresh graduates alike, need to develop their expertise instead of an over-reliance on AI guidance and assistance, which will result in an output that is likely not of any use and create results that are unsatisfactory to employers and clients alike.

Institutions are positioned to train and teach students how to use AI tools with accountability, in conjunction with their own personal expertise and knowledge. In essence higher education institutions must play their part in preventing the birth of more “AI Idiots”.

‍

What can be done?

Institutions need to provide advice and clear guidelines on AI usage, for faculty and students alike, touching on the use of AI in assessing students, doing assignments and creating research, with a focus on using AI tools responsibly and wisely.

Setting clear and extensive institutional guidelines and standards for AI use will help provide fairness and clarity for faculty and students on how they can integrate AI resources into their learning and teaching. Institutions can also introduce courses focussed on AI that can provide students with a foundational understanding of how AI is developed, works and can be applied.

By creating a regulated environment in education where students can learn how best to make use of AI tools, while developing their own skill sets and knowledge, institutions can play a key role in helping future generations harness AI efficiently and effectively.