The primary focus in any discussion about AI in recent years has ultimately come down to the subject of productivity. This narrative has been utilized since the early days of the technology industry, starting from the boom that came with Windows 95 to the present: mention productivity, and you’re in. Remember those days while I was a senior technology analyst and everyone was hyped by Window 95 as a productivity improving product which had an amazing ROI within the first year? The reality, on the contrary, was that, plagued by various technical issues, it was the cause of very long delays and disruptions, and in turn was detrimental to productivity.
Now that we are far ahead, the same is also happening with AI: AI is being touted as something that will change industries or increase productivity, but actually, the very pressing issue here is the quality of the decisions we make. This century’s dilemma is ironically not productivity nor efficiency but quite simply the lack of decision support. How fast the next tech improvement happens is pure joy, true, but are those the appropriate decisions we are making?
I noted a familiar aspect to this while attending Computex for prep not too long ago. The slideshow reports revealed big talk about the AI-enabled acceleration of time, yet everyone stopped short of elaborating on why this change would inevitably translate more effectively to better decision-making. I worry that only an emphasis on speed and ignoring improvement of our decisions themselves might lead us to make machines to err at unprecedented speeds, mistakes that might never be possible to correct.
Productivity versus quality: the age-old dilemma.
During my years in IBM, I learned how critically important quality was, perhaps even more than any other dimension. I vividly remember being in a class of the Society of Competitive Intelligence Professionals (SCIP), where the instructor taught one principle: “Speed without direction is dangerous.” The reasoning is simple: if one keeps speed as a criterion ahead of direction, one could end up running furiously down the wrong track. Speed, therefore, is not a criterion for success; in fact, it could spell calamity when associated with failure.
During the course of my time in IBM and Siemens, I frequently found myself involved in decision support, making well-researched recommendations to decisions that were either ignored, if not actively campaigned against. The ensuing consequences? Catastrophic losses and failure. I learned that many executives would stick to their gut feelings rather than accept data-based advice for fear that to do so might hurt their reputation and career. The overriding of reality with appearance resulting from this mindset has, of course, led to the failure of many initiatives and, in some cases, indeed disbanded departments, including mine.
When I assumed the role of an external analyst, I found that the executives listened to my recommendations more. I was not, in their eyes, a competitor to their careers since I was not involved with the company politics. It reinforced for me an important lesson: AI should enhance decision-making, not just serve productivity and performance. Decisions that lead speed to be worthless if bad decisions are behind it. A fast wrong decision is still wrong but carries with it an earthly amount of collateral damage.
The Artificial Intelligence Dilemma When It Comes to Decision-Making
AI certainly has the capacity to make decisions at a speed greater than that of humans, but more and more questions are being raised around the quality of such decisions. And when we use companies like Microsoft and Intel, which could be seen as leading lights among so-called AI companies, it is practically obvious. Both have, within the last couple of years, gone through upheavals in leadership owing to the poor decisions made in those companies. A striking example in that sense is Steve Ballmer, former CEO of the software giant Microsoft and perhaps the most cited case of a poor decision-maker with a wealth of information at his disposal. During his term, Zune was rolled out, Microsoft Phone launched, and Yahoo was acquired—all of them decisions that curbed Microsoft growth and prompted his exit.
I had the experience of working directly with Ballmer and many others in the company’s executive ranks and advising them on making decisions a bit better. It never really happened. Most of my suggestions were ignored. It was really unfortunate to witness how often corporate executives would be willing to side with their egos and those of their buddies in matters of advice above outside valuable input. But such homespun philosophy isn’t limited to the company Microsoft. Such was an IBM experience under CEO John Akers when all the available internal intelligence was blocked out by executive teams who preferred to trust their instincts and those of their chosen few reliable allies.
This is pretty typical of AI: even though the technology is impressive, by current quality concerns, it still fails to reach its true potential in the area of decision support. AI will only improve, without changing that characteristic reliance of executives, but they again will be taking gut feelings instead of what the data from AI offers. Trust grows down into this crevice of systems that won’t or can’t be trusted and continues the cycles of outdated behaviors in decision-making.
High Speed Is Not Everything-AI Must Focus on Quality
There is a fundamental thing that is being ignored by productivity and performance. We live in a time when AI improves the speed and productivity of humans but at the same time delivers reliable, accurate and actionable results. Quality counts.
Until that trust issue has been dealt with- that is, the individual is confident enough in AI’s recommendations to make use of them- it’s hard to see how the technology will fulfill its promise. Argumentative Theory states that advice received from inside may be construed as a threat to one one’s rank or job. Hence, brilliant pieces of advice become rejected when they’re subject to possible future question by AI by executives.
Thus, decision support delivered from AI should improve first. Only when we can evaluate trustworthiness of decision support should we even move on to productivity. A speedier misjudgment is still a final error and AI makes sure it directs rather than speed alone.
The Road Ahead: Training for Effectively Using AI
AI can add value only when leaders and organizations are trained to embrace the insights AI can offer. This requires a culture shift in the corporate world where people will be rewarded for choosing actions based on AI output and not penalized for trusting technology. Only an ambiance in which people perceive AI as a mild instrument in decision-making-no more-to avoid personal loss will bring about the greatest potential of AI.
AI can change the way we work, how decisions are made, and how we shape different industries. It is not just with speed that we kill the snake, but rather with IT, to have a source from where one can hope to get delight. Quality, ethics, and consulting as decision-making assistance are absolute must requirements in this respect, ensuring that AI assists in navigating through the complexities that lie ahead. This must stop from being seen as a tool for increasing productivity, and now the time has come for AI to make the best and most informed decisions for our benefit.
Product of the Week: Starlux Airlines
And on something a bit less heavy, I must also report my delight for the service rendered by Starlux Airlines, a product of Taiwan. A consistently disappointed customer with United Airlines, I was blown away with the improved service Starlux had to offer. Attendants from the airline went out of their ways to take care of various special requests, such as specific meal choices for dietary needs and even a passenger who needed help in accessing the Wi-Fi. It was refreshing to be so genuinely cared for on a flight. My experience reminded me of how non-US carriers, such as Starlux, Singapore Airlines, and Emirates, are always customer-centred, compared with their American counterparts. Starlux is one sure company that has won my respect and can thus hold the quality as a choice of this week.