How to adopt AI in a way that makes sense
The near-blind adoption of AI doesn’t make much sense. It may even hold your projects back if the technology is not implemented properly — which is often the case.
Ayanna Howard, an executive counselor on the board of the Association for the Advancement of Artificial Intelligence (AAAI), has a front seat when it comes to observing the AI practices adopted by companies. Howard identifies three recurring problems:
1) When taking AI on board, the “why” and the “how” are not properly considered. AI is not an end in itself; it’s an answer to one or more clearly identified problems.
2) AI’s value is based on the data you feed into it — but data is too often erroneous (or worse). In the absence of “why,” valuable data can easily lead to decision-making mistakes and even grave offenses (remember Tay, Microsoft’s chat bot and its inflammatory tweets?).
3) Businesses do not always balance the value AI offers with the personal privacy and agency that is lost.
In a nutshell, Howard argues that decision–makers do not ask enough questions. She offers two recommendations: (1) AI requires the right skills to understand it, so either an expert internal or external team must be available; and (2) leaders must ask the right questions and develop AI that is truly useful, by consulting all stakeholders and establishing clear ethical standards.
To go further: “Demystifying the Intelligence of AI” by Ayanna Howard (MIT Sloan Management Review, November 2019)