AI and Ethics: The Waters are Murky, but Help is Available

R. Bhattacharyya

Summary Bullets:

• Many organizations need help navigating ethical issues related to artificial intelligence (AI), such as privacy laws, unintentional bias, and lack of model transparency, but don’t know where to begin.

• Enterprises can benefit from working with a partner that helps them consider the ethical implications of their AI deployments, but they should keep in mind that issues aren’t static and can evolve over time.

Organizations are eager to enjoy the benefits that AI can bring to them – whether enhanced productivity, or new revenue-generating or enhanced customer experience opportunities. But many are unclear about how to navigate the murky waters of AI and ethics. Changing regulations and privacy laws, concerns over unintentional bias in training data, lack of transparency in AI models, and the dearth of experience with new use cases are difficult challenges to address. Enterprises want to ensure that their adoption of the technology doesn’t cross ethical boundaries, but often don’t know where to begin. Thankfully, the topic is being increasingly addressed by IT services providers. Many organizations, from IBM to Capgemini to Atos are touting that they help their customers implement AI while also considering the ethical implications of their deployment.

France-based Atos has established itself as a thought leader in the space by developing an ‘Ethics by Design’ framework to help its customers navigate what is generally unfamiliar terrain for most organizations. The company has an Ethical Advisory Board and promotes a multidisciplinary approach that encompasses tools, methods, governance, culture, and regulation. Components range from bias detection, model traceability, and analysis of the impact of regulatory changes, to use case analysis and definition of corporate values. Taking a different approach, IBM has launched AI Fairness 360, a toolkit that helps developers and data scientists examine, report, and mitigate discrimination and bias at multiple stages during machine learning model development. It contains more than 70 fairness metrics and ten bias mitigation algorithms.

Enterprises would be wise to work with a partner that prioritizes ethical considerations and has compiled best practices from numerous client engagements. But ethical issues aren’t static; they can evolve over time. A good partner will continually revisit its ethics framework to ensure it reflects the latest regulatory changes, technology trends, and user preferences. Furthermore, ethics reviews shouldn’t be a one-time event conducted during solution deployment; issues related to AI and ethics need to be constantly reviewed. The topic of AI and ethics is a fluid one, with issues likely to change as the technology matures and societal concerns and preferences evolve.



What do you think?

Please log in using one of these methods to post your comment: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.