Why AI requires ethical consideration

Share this post:

Author: Ross Farrelly, Segment Leader, Data Science and Artificial Intelligence IBM Global Markets – Cloud Sales

Business leaders agree that AI should be used ethically, and many believe in its potential to do good both in business and in broader society. But how do we ensure that our AI is making ethical decisions on our behalf, and whose responsibility is it to ensure this?

A study by the IBM Institute for Business Value reported that more than half of the 1,250 executives surveyed believe AI can improve their companies’ ethical decisions. Yet while 8 out of 10 directors surveyed believe the ethical questions raised by the deployment of AI are board-level issues, only 45 per cent feel fully prepared to oversee them.

AI needs to be transparent and trustworthy

With AI adoption rapidly increasing, there is a need to translate the abstract theory of ethical AI into practical guidelines. And as AI is continuously evolving and being applied in more and more mission-critical situations, there is a growing need for the decisions made by AI systems to be transparent, explainable and trustworthy.

A small error in AI output might cause irritation in an everyday scenario but could have serious ramifications in a business context. For example, on social media, if an AI system wrongly labels a face when grouping photos, it’s a minor inconvenience. But if an AI system makes a biased judgment based on personal data and an insurance company then incorrectly approves or rejects a claim or if a bank incorrectly approves or rejects a credit application based on that judgment, the consequences can be significantly more severe.

Scenarios such as these highlights the need for both data sources and predictive models to be scrutinised at every step of the ladder to AI. Processing large volumes of semi-structured data is just one way in which AI can assist people to make faster, more informed decisions. However, AI is only as good as the data it is fed, and biased data and can generate biased results.

A question of commitment

AI has the potential to do good in businesses by saving time, energy and cost, and it has untapped potential to do good in society in general. Smart, rapid and trustworthy decisions are the aspiration of companies which are employing AI, but the journey to achieve this requires investment and ongoing learning to develop a secure, scalable, trustworthy AI infrastructure.

Ensuring that AI responds and acts ethically is a question of commitment. Ethical AI can be achieved by building a trusted AI infrastructure and commitment to ongoing improvement. While leaders won’t know the immediate answer to every scenario, it is still their responsibility to ensure that their AI responds appropriately and ethically.

If you’d like to expand your knowledge of AI and what’s required for AI success, you can watch our on-demand content for the Data & AI Virtual Forum ANZ here. With keynotes from industry leaders and deep-dive technical sessions, we trust that these resources will leave you ready to unleash the potential of AI in your business.