Latest News

Principles and Practices for Building More Trustworthy AI
October 06, 2021

By Seth Dobrin, Chief AI Officer and Christina Montgomery, Chief Privacy Officer & AI Ethics Board Co-Chair AI is transforming almost every aspect of how businesses operate and engage with the...

By Seth Dobrin, Chief AI Officer and Christina Montgomery, Chief Privacy Officer & AI Ethics Board Co-Chair

AI is transforming almost every aspect of how businesses operate and engage with the world. This technology can unlock the potential of data and revolutionize our daily routines by automating simple and repetitive tasks. AI will be key to help address massive global challenges like pandemics by accelerating drug discoveries and more. At the same time, there are also valid concerns about how this technology is being used, and governments are starting to respond. In April 2021 for example, the European Union unveiled its proposed regulation designed to address the potential threats to health, safety and fundamental rights posed by AI systems. As governments explore the critical questions presented by the advancement of AI, companies also must take responsibility for how their own organizations build and use AI.

A new World Economic Forum (WEF) case study lays out the stakes and affirms IBM’s leadership in trustworthy AI. Used irresponsibly, AI has the potential to erode trust, propagate inequality and create harm. Organizations that want to use AI have a fundamental responsibility to foster trust in AI solutions. IBM has been guided by this view for years, and our leadership in this space resulted in IBM's 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.

To dive deeper, here are some of the practices IBM has developed which help us build AI capable of fostering that trust.

Ethical Principles at the Core

Because developing and managing trustworthy AI is an ongoing process, IBM applies high-level Principles for Trust and Transparency to continually build and strengthen that trust. The principles 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.

IBM also embeds ethical principles across our company's global operations through an AI Ethics Board which provides centralized governance and decision-making authority. The AI Ethics Board helps instill a culture of technology ethics throughout the company and is one mechanism by which IBM holds our company and all IBMers accountable to our values and commitments for the ethical development and deployment of technology.

Governed Data and AI Technology

It’s easy to say that ethics matter, but actually embedding those ethical principles into the technology itself is more complex. Organizations recognize the importance of a holistic approach to managing and governing their AI solutions across the full AI lifecycle. Our goal is to bring together products, services, systems, and research assets to develop solutions specifically designed to help businesses not only set and execute AI strategies, but to help infuse trust into their existing and future AI systems and put them into operation.

Many of IBM’s innovations in trustworthy AI are born in IBM Research and built on our five focus areas: explainability, fairness, robustness, transparency and privacy. Our IBM Watson solutions and consultants from IBM Global Business Services help businesses with auditing and mitigating risk, implementing governance frameworks, operationalizing AI, education and guidance, and organizational change. Organizations ranging from large American retailers to financial institutions like Regions Bank and sports organizations like ESPN Fantasy Football are putting the principles of trustworthy AI to work.

Our AI FactSheets advance trust and transparency even further. AI FactSheets make AI more explainable by listing the components, goals and information about how it works and was developed, not unlike a nutrition label, and we’ll soon be making these AI Factsheets part of IBM Cloud Pak for Data.

An Open and Diverse Ecosystem

We believe that AI should benefit the many, not just an elite few. Delivering on that means fostering a culture where diversity, inclusion and shared responsibility are imperative. This includes a diversity of datasets, diversity in practitioners, and a diverse partner ecosystem to enable continuous feedback and improvement.

What’s Next

AI’s benefits stand to grow exponentially, but this can only happen if society trusts it. That’s why we believe AI systems must prioritize consumer privacy and data rights. It’s also why we have called for the precision regulation of AI to place the tightest regulatory and policy controls on technology end uses where the risk of societal harm is greatest. And we’re not waiting for regulation to take effect before getting this right.

With trust as the cornerstone of our leadership in AI innovation, IBM is the partner that businesses need right now as they look to use AI as a force for positive change. And at this moment in the long arc of human progress, that matters not just for our company, but for our customers and society at large.

If you are interested in reading more, check out the new World Economic Form case study that explores the evolution of IBM’s approach to trustworthy AI in greater detail, the IBM Watson Trustworthy AI page and the IBM AI Ethics page. You can also explore some of the ways our technology and consulting services are helping clients here, review the IBM Institute for Business Value research on the importance of AI ethics among boards and executives and sign up to join the list for our upcoming AI FactSheet open beta which we expect will be coming to IBM Cloud Pak for Data later in October.

Article Categories