LONDON - Artificial intelligence (AI) and machine learning are rapidly expanding their reach into numerous aspects of people’s lives. These technologies are even finding their way into socially sensitive domains such as labor markets, education, criminal justice, healthcare, banking, and housing. Given their ubiquitous use in information curation on search engines, social networks, and recommender systems as well, it is of the utmost importance that the ethical implications of their deployment be addressed.

The increasingly heated debates on ethics in AI reflect the need for society to forge a consensus regarding the appropriate directions for technical innovation, the responsible use of these technologies, and the constraints that should be imposed upon them. This task is far from simple but is essential if humanity is to shift the focus from distracting questions of AI outperforming humans and ‘taking over the world’ to a more constructive discourse.

This article will examine the state of play regarding reliability and trustworthiness, map out the current debates surrounding these issues, and examine the human-machine relationship. It will also explore how the dichotomy of ‘human versus machine’ pushed by science fiction and mainstream media is holding humanity back and provide some suggestions about how people might foster greater collaboration and shared understanding.

AI and the world today

Safety and trustworthiness are now key considerations in two strategic areas of AI: embedded AI in physical systems - such as aircraft and trains - and general-purpose AI, such as in natural language processing and task completion.

Despite the frenzy about the growing power and smartness of these tools, it is important to bear in mind that they are far from perfect - the gap between the development of AI tools and their actual adop