Three Steps to Embracing Artificial Intelligence – CDOTrends

Despite its strategic capabilities and potential to significantly advance how we do work, the average user is understandably wary of AI. Foremost among their concerns would be a fear that AI could take away their jobs outright,  or at a minimum render them to redundant over time.But many of these concerns are not necessarily based on what is happening in the field of AI. Writing on the World Economic Forum Bowen Zhou, the President of Cloud & AI chairman at noted that much of these fears stem from a less than complete understanding of AI.He offered three tips to address unrealistic fears over AI and to bring about success.Explain AI moreWhile many people imagine AI as an extremely smart machine that can make all the decisions, the truth couldn’t be further from that. Indeed, the current developments in AI often entail large amount of data, and competencies in extremely narrow fields such as language or vision processing.Instead of writing off AI as being too complicated for the layperson, Zhou suggested making a concerted effort to ensure that society can be understood and accessed by everyone. This will need to happen through education, as well as collaboration with AI scientists and engineers to ensure that people understand the potential and limitations of AI in the community.Bring everyone alongThere is no question that we are but at the very nascent stage of AI. Despite the seemingly large advances in certain niche areas, there is certainly a lot more to be explored in this field.For instance, more than half of current AI funding are centered around machine learning applications and platforms, with smart robots, computer vision platforms and natural language processing getting most of the rest – and it tapers off significantly after that.Zhou pointed to possibilities such as unsupervised learning (or transfer learning), or for AI to learn in a much more natural way. Given that we are only just starting, he suggests that the next breakthroughs in AI can come much faster with the participation and engagement of the global community working together.Agility with governanceFinally, governance is a vital topic that needs to be addressed. While Zhou doesn’t deny that, he observed that it took 60 years for the U.S. to issue formal regulations on the use of seatbelts after Ford Motor produced the Model T car. The argument? Don’t prevent the global community from enjoying the use of AI just because we are still working on regulations around AI.Obviously, it helps that organizations adopt the use of AI in a responsible way, by taking ethical guidelines into consideration. For this, Zhou points to the six dimensions of a “Trustworthy AI” which has been posited by some in the international community. These typically revolves around fairness, reproducibility and explainability.Writing separately for Forbes, Naveen Joshi outlines some of these guidelines for developing a trustworthy AI in more detail.Explainability: Organisations that develop AI must open the “black box” to understand how results and conclusion and arrived at.Integrity: Output should be generated in accordance to a developer’s predefined parameters, both operational and technical ones.Conscious development: Decisions by AI systems must benefit the human raceReproducibility: Every outcome generated by an AI system must be reproducible.Regulations: Government organizations must develop guidelines and regulations.For now, there is no question that AI is slowly but steadily making its way into traditional businesses, including a floating fish farm in Singapore. You can read more about it here.Photo credit: iStockphoto/metamorworks