Can AI ever be “ethical”?
Ethical AI is becoming a buzzword in industries as more use cases of AI are being developed globally. What was once taunted as an expensive piece of technology is now available in most industries. Like every new piece of technology, AI has its limitations as well. And in this case, it’s whether the technology is ethical or not.
Despite that, the technology can learn and harness data faster, providing better results in the use cases it’s being implemented in. Be it taking photographs, driving a car, ordering food, and doing transactions online, there are elements of AI working in the background all the time.
For enterprises, AI has enabled them to not only improve their productivity but also have better analytical insights and also make data-driven decisions. Together with cloud computing, businesses are deploying AI to achieve IR4.0.
However, with greater technologies, comes greater concerns as well. With AI technology still growing and learning, there are concerns about the ethics that come with AI. As AI feeds primarily on the data it has access to, there have been issues on AI potentially eroding trust, exacerbating inequity, and causing harm to both people and the environment.
While it is not a case of machines trying to take over the world, some AI algorithms have been known to be biased and being unethical. As such, some consumers are having less trust in the tech industry despite the increasing adoption of AI in most business use cases.
The World Economic Forum and the Markkula Center for Applied Ethics at Santa Clara University published a case study highlighting processes, tools, and organizational constructs that facilitate the responsible design, development, and implementation of technology. The new white paper recognizes IBM’s leadership as a trusted, responsible technology company, particularly as a leading developer of ethical AI.
To understand more about ethical AI and its impact on both consumers and businesses, Tech Wire Asia spoke to Seth Dobrin, Chief AI Officer, IBM.
In the first part of this interview, Seth explains what ethical AI is all about and how IBM is implementing it in its solutions for use cases around the world. Seth also shares how blockchain and ethical AI are linked.
What exactly is ethical AI?
AI ethics are an essential part of building an AI that is trustworthy, but they are just one piece of the puzzle needed to help people and organizations adopt AI responsibly. We focus on a human-centered approach to trustworthy AI, an approach that puts ethical principles at the core of our governed data and AI technology and fosters an open and diverse ecosystem.
When it comes to our ethical principles, our guiding values make clear that the purpose of AI is to augment human intelligence; that the data and insights generated from data belong to their creator; and that powerful new technologies like AI must be transparent, explainable, and mitigate against harmful and inappropriate bias.
It’s easy to say that ethics matter, but actually embedding those ethical principles into the technology itself is more complex. That’s why the second pillar to our approach is to continually bring innovative governed data and AI technology and approaches to market that is built on five focus areas: explainability, fairness, robustness, transparency, and privacy.
Finally, we believe that for AI to be successful, it must be built in an open and diverse ecosystem. Delivering on that means fostering a culture where diversity, inclusion, and shared responsibility are imperative.
AI is already transforming how businesses operate and engage the world, delivering the power of prediction to augment human decision-making. However, humans must be able to trust predictive recommendations and outcomes for AI to realize its full potential. That’s why IBM believes it’s essential to consider AI ethics as one key component of the holistic approach that’s required to build trustworthy AI.
How do vendors define it in their solutions, especially when dealing with customers?
Organizations today recognize that it takes a holistic approach to manage and govern their AI solutions across the full AI lifecycle. There are good evidence companies and their customers are on the same page when it comes to the need to achieve fairness and reduce bias. A recent survey showed 91% of businesses using AI today say their ability to explain how it arrived at a decision is critical and 86% agree that consumers are more likely to choose the services of a company that is transparent and ethical about how the AI they use is created and managed.
That said, while the data shows global businesses are now much more aware of the importance of having trustworthy AI, more than half of companies cite significant barriers in getting there including lack of skills, inflexible governance tools, biased data, and more. It’s clear while there are tools and frameworks in the market to help build trustworthy AI, there is still work to be done to help businesses develop a comprehensive approach to AI governance.
That’s our current focus at IBM. We’re working to continually bring innovative governed data and AI technology and approaches to market that are built on five focus areas: explainability, fairness, robustness, transparency, and privacy. Our solutions for trustworthy AI help businesses with things like auditing and mitigating risk, implementing governance frameworks, operationalizing AI, education and guidance, and organizational change. Organizations ranging from a large American retailer to financial institutions like Regions Bank and sports organizations like ESPN Fantasy Football are putting the principles of trustworthy AI to work.
Is there a link between blockchain and ethical AI?
Blockchain is a shared, immutable ledger that facilitates the process of recording transactions and tracking assets in a business network. Almost anything of value can be tracked and traded on a blockchain network. Blockchain is ideal for delivering that information because it provides immediate, shared and completely transparent information stored on an immutable ledger that can be accessed only by permissioned network members.
AI, until recently, was not built with issues like security or transparency as primary goals. But, at IBM, we see these components as central to building trust in the technology. That’s why our approach to ethical and trustworthy AI is built on our five focus areas: explainability, fairness, robustness, transparency, and privacy.
That said, in the second part of our article, Seth explains how ethical AI deals with AI biasness in some use cases and how businesses can deal with them.