
When it comes to the ethics of artificial intelligence, technology leader IBM takes a ‘multidisciplinary, multidimensional approach to trustworthy AI’. IBM was recently recognized for the third year in a row as one of the World’s Most Ethical Companies. This achievement was given to a total of 135 companies across 22 countries and 47 industries. Ethisphere honors companies each year based on culture, environmental and social practices, ethics and compliance activities.
IBM is a 110-year-old company that has adapted its business values to changing ways of life. For example, IBM recently established their AI Ethics Board. Guided by principles of trust and transparency, IBM pursues ethical data practices with a trickle down policy. Following their principles, IBM’s foundational properties for AI ethics include explainability, fairness, robustness, transparency and privacy. The Ethics Board then comes into play to put these principles into action across the organization.
The IBM AI Ethics Board regularly addresses these positions:
- Precision regulation for AI
- Racial justice reform
- Data responsibility
- Facial Recognition
- Mitigating bias in AI
- Learning to trust AI systems
AI is systematic in our day-to-day lives and provides behind-the-scenes improvements for millions of people who may not even realize it. Innovations in agriculture, healthcare, supply chains and car maintenance are just a few examples of artificial intelligence’s reach. Even with all of this good, there are still questions of the negative impacts and ethical implications of this advanced technology. IBM addresses these issues and more through their Ethics Board.
Beginning with facial recognition, IBM recognizes that our faces are the most intimate and personal expressions of our identity. It is reasonable to be uneasy about technology that can capture and remember individualized features. On the other hand, IBM reinforces that blanket bans on this technology introduce bigger risks, such as cutting off society from the many benefits of facial recognition – some of which can be life saving.
Such blanket bans on facial recognition technology overlook the distinct categories of these systems. Facial recognition technology is broken into facial detection, authentication and lastly facial matching. Each category raises different societal solutions and according concerns. IBM Ethics Board recommends that companies and policy makers adopt a precision regulation policy that applies restriction and oversight to particular instances instead of applying umbrella policies. “It simply does not make sense to subject a smartphone and a police body camera to the same regulatory treatment. The same technology used in different situations by different users should be governed by different rules.”
The policies and practices of the IBM Ethics Board are simple and provide great clarity, they make it seem easy to address data ethics by differentiating between issues and making reasonable statements that seem to be lacking in many conversations of data ethics today. As a general rule, IBM states that organizations should obtain consent before using facial recognition. They also recognize more obvious limitations–for example the use of facial recognition for mass surveillance or racial profiling is a clear violation of basic human rights and freedom, and providers of this technology must be accountable for assuring their software is not facilitated in abusive ways.
IBM acknowledges that the best way to devillianize AI and advanced technology is to provide complete transparency and unassuming explanations to the general public. Their policies are informative and action-based in order to promote positive AI practices and acceptance for the usefulness of this technology in every person’s daily lives.