Editor's note: On June 13th, Glenn Gow will be one of the keynote speakers at the next AI Unleashed event. This event is a concise, half-day online forum designed to help board members quickly understand their responsibilities for risk oversight of generated AI. It can help you better guide strategic conversations with management. Why not join us >
Since the rapid emergence of ChatGPT and its artificial intelligence, I've answered many questions about how it will impact the way we work. Senior executives and board members understand that generative AI will change the way we work, but the devil really is in the details.they want to know which one work? how Will they change?and what What leaders can do to get ahead of chaos.
To answer these questions, we find it helpful to refer to Singapore's Model AI Governance Framework (AI Governance Framework). The Singapore Model was developed to help organizations deploy AI responsibly and effectively, and considers the deployment of AI relative to the role of humans in its use. (See image below)
A key aspect of this form of AI governance is its focus on “harm.”
1. The severity of the harm that AI may cause in a particular role (e.g., economic harm, reputational risk, public safety issues).
2. The potential for this harm to occur if AI is left alone without human involvement.
Therefore, AI governance frameworks provide very useful models for assessing which tasks are suitable for increasing AI autonomy and to what extent human authority in decision-making should be retained. Offers. Let's take a closer look at three possible scenarios.
Human in the loop: (Quadrant 2 — top right)
explanation: Although AI can augment and support human decision-making, humans are still essential for execution due to the high potential and severity of AI's harm.
example: Examples of roles in this quadrant include medical interpretation (e.g., radiology), software development, legal analysis, drug testing, complex customer service problem solving, human resources (hiring and performance evaluations), corporate finance, and due diligence. included.
Leadership tips: For roles in this quadrant, AI can significantly reduce the most tedious and time-consuming tasks. This significantly increases human productivity, increases job satisfaction, and improves both the quantity and quality of work products. To ensure success, the administrator must choose wisely the tasks that her AI will undertake. They also need to educate and guide their staff to understand the “auxiliary role” of AI so that human team members still feel a strong sense of ownership and responsibility. (See article: What can you do with all that productivity?)
Human crossing the loop: (quadrants 1 and 4 – top left and bottom right)
explanation: Although AI can operate without humans, good judgment encourages human oversight because the potential or severity of harm is too great to allow AI free reign.
example: Jobs in this quadrant include autonomous transportation, medical and dental treatment planning, sales and marketing functions including design and written communications, stock/commodity/other asset trading, predictive industrial maintenance, and all types of advisory services, e.g. financial planning, employee benefits, job scheduling) and financial management.
Leadership tips: The superior capabilities of AI in this quadrant often allow it to operate without continuous human input. However, due to the likelihood and severity of the potential harm that can be done in these areas, humans must supervise this work and intervene when human discretion is necessary. Managers should study the risks of using AI in these scenarios, pay close attention to evolving laws governing the use of AI, and implement strong training programs for human supervisors. Supervisors also need to be wary of complacency in the face of AI that works perfectly “almost always.” (See article: AI risk management for corporate boards)
Humans outside the loop: (Quadrant 3—bottom left)
explanation: The likelihood or severity of harm resulting from the use of AI is so low that AI can make decisions as good or better than humans and can operate autonomously.
example: Jobs in this quadrant include product recommendation systems in e-commerce, weather forecasting, supply chain demand forecasting, advertising media buying, traffic flow forecasting, and industrial or commercial building environment management.
Leadership tips: Humans are out of the loop, so little training or special supervision is required to perform this task. However, it is important to set up effective performance monitoring as part of a feedback loop between affected staff, the market, and software developers. This area of AI promises the highest productivity at the lowest cost for organizations, but the business environment is constantly evolving, so “set it and forget it” is not really the case.
That's it! Considering the severity and probability of harm that can be caused in a particular work area provides organizational leaders with an effective filter for determining “how much.” And “How soon?” It is also possible to introduce AI. This helps leaders create a roadmap for implementation, starting with the areas where both the likelihood and severity of harm is lowest.
