NetDragon Websoft, a Hong Kong-based online gaming company boasting an annual revenue of $2.1B, took an unconventional step in August last year. They appointed a new CEO for their primary subsidiary.
The new executive, Tang Yu, handled all the standard responsibilities of a corporate leader: analyzing high-level metrics, making strategic decisions, assessing potential risks, and promoting a productive work environment.
Yu was an indefatigable worker, operational 24/7, never needing rest, and accepting no annual salary. Quite a commitment, indeed.
Under Yu’s leadership, the company has outperformed the Hong Kong stock market since her appointment.
However, there’s an unexpected twist: Tang Yu was no ordinary CEO. Actually, she wasn’t a human at all. She was an AI-powered virtual robot.
Conventionally, automation efforts have been focused on so-called lower-tier and blue-collar positions.
More recently, AI has begun to encroach on white-collar professions like accountants and journalists.
Interestingly, while the executives at the top of the corporate hierarchy laud the cost-saving benefits of AI, they seldom consider applying the same logic to their own positions.
The motivation for workplace automation is primarily financial. So, why not begin with the most expensive employee on the payroll – the CEO?
Who knows, perhaps one day employees might be secretly wishing for a software update that introduces casual Fridays.
As the capabilities of AI have expanded, automation experts have pondered the possibility of replacing a significant portion of the workforce.
A recent prediction by the McKinsey Global Institute suggests that by 2030, about 45 million workers, or approximately 28% of the entire American workforce, could lose their jobs to automation.
Historically, automation has targeted roles typically categorized as lower-level and blue-collar jobs such as warehouse employees, truck drivers, office assistants, and food preparation workers. However, the advent of more advanced AI has now begun to threaten white-collar professions such as accountants and journalists.
The primary motivator for workplace automation is financial, so why not consider replacing the highest-paid employee — the CEO?
The Economic Argument for Automating CEO Positions
In Fortune 500 companies, the average annual CEO pay is now roughly $16 million.
Over the last 45 years:
- CEO pay has increased by 1,460%
- Average worker pay has only increased by 18%
Consequently, the average CEO today earns the equivalent of 399 median workers.
For larger companies, this ratio is often even higher.
For example, in 2021, Amazon CEO Andy Jassey received a package worth $213 million — the equivalent of the combined salaries of 6,474 Amazon employees. This workforce could fully staff four fulfillment centers. A significant part of a CEO’s role is to act as the face of the company — a task fairly resistant to AI.
However, this doesn’t imply that many other aspects of a CEO’s role can’t be automated.
McKinsey estimates that about 25% of a CEO’s time is spent on tasks that could potentially be automated by machines or AI — reviewing financial performance, sending emails, predicting trends. Another possible area for automation is the executive decision-making process.
Annually, executives make approximately 3 billion decisions — the effectiveness of these decisions is directly tied to a company’s financial performance. Hence, there’s an incentive to improve the success rate as much as possible.
CEOs are often lauded for their decision-making prowess. However, by their own admission, they’re not significantly better at making good decisions than anyone else:
- A McKinsey report revealed that 72% of executives confessed to making bad decisions as often as they made good ones.
- At the typical S&P 500 company, ineffective decision-making results in losses and missed opportunities amounting to $250 million per year.
To improve these ratios, executives are increasingly turning to algorithms and machine learning for assistance. But what does AI think about this potential shift?
We consulted ChatGPT, the powerful AI chatbot that’s reigniting concerns about automation, to ponder the probability of a widespread CEO replacement.
Here’s the response:
“AI is not poised to replace the job of a CEO in the foreseeable future. The job of a CEO often involves highly complex tasks such as setting strategy, inspiring and leading employees, building relationships with various stakeholders (e.g., investors, partners, government officials), and making complex, nuanced decisions that often require a deep understanding of the human element. While AI has made significant strides, it is still not capable of fully understanding and replicating these types of human cognitive and social capabilities.
AI can, however, assist CEOs in various ways. For example, AI can help analyze large amounts of data to provide insights for decision-making, automate routine tasks, and predict future trends.
In the future, it’s possible that AI might take on more roles within organizations, but this will likely be in a supporting role rather than completely replacing human leadership. The human element in leadership—such as empathy, intuition, and ethical judgment—is crucial and currently irreplaceable by AI.
However, it’s worth noting that the role of AI in businesses and society is a rapidly evolving field, and it’s hard to predict with certainty what will happen beyond my knowledge cutoff. As AI advances, it’s important for societies to have conversations about the ethical implications and the potential impact on jobs and the economy.”
One notable voice believes that time might be running out for CEOs.
Years ago, Jack Ma, the Chinese billionaire co-founder of Alibaba Group, speculated that the unemotional logic and efficiency of AI could eventually make its way to the executive suite.
He predicted that in 30 years, “a robot will very likely be on the cover of Time magazine as the best CEO.”
Given the performance of NetDragon’s AI-powered CEO, who continues to outperform the stock market, Ma’s prediction might not be as far-fetched as it seems.