IBM Watson

In Case You Missed it: Francesca Rossi, IBM's Global Leader for AI Ethics on the Challenges of Governing AI

July 13, 2018

Francesca Rossi, IBM’s global leader for AI ethics, recently spoke at "Governing Artificial Intelligence," a conference organized by the International Peace Institute and the United Nations University's Centre for Policy Research.

The panel explored the unique ethical challenges posed by the recent and rapid advances in AI. Here are some of the most important takeaways from Francesca Rossi's intervention:

- Human + Machine: The purpose of AI is augment professionals in everything they do. But for human and machines to work in unison, there must be trust in the technologies we usher into the world. IBM is working hard to create the right level of trust between men and machines.

- Algorithmic bias: Algorithmic or human bias can seep into AI models. IBM is providing engineers with the intellectual and technological tools they need to keep AI bias in check.

- Consumer AI vs Business AI: IBM's Watson is an AI platform specifically designed for the enterprise. In terms of ethical challenges, there's a clear difference betweens AI that help doctors identify treatment options and an AI system that recommends books.

- Explainable AI: For AI to flourish, AI must be able to explain itself. This means that companies advancing or using AI must provide explanations for how an AI system came to a particular decision in a way that is easy for people to understand.

- Diversity and inclusion: The development of AI needs to be underpinned by a spirit of diversity and inclusion. By giving voice to multiple points of view, we can protect our algorithm from becoming an echo chamber.

- Data responsibility: For businesses to reap the benefits of AI, companies have a duty to handle data responsibly. They must be clear about who owns that data and who are they sharing it with. IBM's own trust and transparency principles offer a roadmap. In it, IBM is clear: data and the insights they generate belong to their creator.

Francesca Rossi concluded her remarks by highlighting the industry's collective responsibility in not only creating AI systems that are trusted, accountable, transparent and explainable, but also in managing the consequences of the AI revolution. This includes the imperative of training and re-skilling for new collar jobs and of which IBM's P-Tech program is a prominent example.