Is AI Overhyped? 5 Leaders Weigh In
Charles Fleet
Operations Executive | Authentic At Scale Team Builder | M&A, Strategy Execution and Continuous Improvement
Four colleagues and I weekly explore how Artificial Intelligence is impacting Leadership as part of an Agile Leadership Journey AI cohort. While we agree that "Yes, it is!!!", we diverge quickly on how, when and why. Realizing there is power in our own perspectives and in exploring each other’s, we are taking on the current topic of “Is AI Overhyped?” by sharing a collection of essays with the larger business community.
Our contributors bring a common base of experience in business agility, leading change, digital business and the pragmatic use of machine learning and AI technologies to the discussion. Yet each writer communicates a distinctive view on our topic.
Leading our group of authors, Len Greski shares the perspective of a c-level executive on why it's important for senior executives to pay attention to AI now, regardless of the level of hype surrounding it. A key reason is that with low-cost subscriptions to generative AI tools, end users can access the benefits of AI without the significant capital investments usually required to train the AI. Expanded access to AI for customers, employees, and competitors creates opportunities and risks that demand immediate executive attention.
Next, Chris Morales addresses the topic from the point of view of a digital marketer, asserting that any short-term disillusionment over unmet AI expectations will be overcome as the industry embeds more of a consumer mindset into AI capabilities and service offerings.
Manoj Vadakkan focuses on potential labor market disruption that may be caused by generative AI from the perspective of a seasoned veteran of multiple waves of disruption caused by advances in technology.
Then Fred Miskawi not only provides a summary of the factors driving the unique trajectory of AI in the marketplace, he also asserts that the industry needs to shift from focusing on AI as a concept to its tangible benefits and outcomes. This shift will enable companies to more effectively generate returns on their investments and avoid being part of the predicted 30% of generative AI projects that will be abandoned after proof-of-concept by 2025.?
Finally, Charlie Fleet anchors our group of essayists with an end user's view on AI's potential impact on our lives so we can become better informed citizens, make better choices about how we as individuals use AI, and guide our companies as they adopt it.?
Please enjoy each leader’s essay in their entirety below.
LEADER 1: Overhyped but Unavoidable: Three reasons why executives need to pay attention to AI now
by Len Greski Senior Technology and eCommerce Executive | Change Management | Digital Transformation | ML / AI | eCommerce Product & Service Development and Management | High Performance Teambuilding
As the old commercials pitched, "it slices, it dices, it Juliennes fries." "It's a floor wax and a dessert topping!" These days the hype around AI, and generative AI in particular, is so pervasive that it's almost impossible to ignore. Many business leaders are intrigued by the outlandish AI sales pitches, but unsure about the ways AI can benefit or harm their organizations. Executives need to become familiar with the development and use cases for AI technologies (especially those that can be directly consumed by end users) for three reasons, including:
Ready or Not…
Although Generative AI was rated at the “peak of inflated expectations” part of the 2023 Gartner Emerging Technologies Hype Cycle, today’s overhyped large language models trace their history at least as far back as 1986, when the work of Geoffrey Hinton and others on backpropagation significantly advanced the training methods for neural networks. Over the next 35 years advances in algorithms, knowledge engineering, storage, and compute technology enabled machine learning algorithms to tackle increasingly complex problems in natural language processing and image recognition, culminating in today’s generative AI products, such as OpenAI’s GPT-3 and DALL-E.
That said, leading companies have been embedding AI capabilities into their products for 35 years, they just haven’t been hyping them as “AI-enabled.” For example, in 1988 I worked on neural network that identified bad grocery store scanner data before it was integrated into a market research firm’s syndicated tracking service. ?At that time, grocery store scanner systems had numerous software defects that caused inaccurate data to be recorded in store-level point of sale systems. ?The major statistics software packages (SAS and SPSS) did not include neural network algorithms, so we had to code our algorithm from scratch.
Subsequent technology advances made it easier for companies to embed AI capabilities into their products and services or to automate labor intensive tasks. Today’s AI-enabled processes range from the use of AI tools to create and monitor cloud infrastructure services to improving the productivity of software developers by 30 – 40%. Furthermore, the availability of free or low-cost Generative AI tools reduces the need for end users to create and maintain the knowledgebases used by the AI tool. Reduction or elimination of this barrier to entry greatly expands the accessible market for AI capabilities.
The bottom line is that now companies can consume AI services through end user interfaces to improve their businesses, so they don’t necessarily need to be experts at developing AI capabilities or the supporting knowledge engineering. If your company isn’t using these services to improve productivity, you will fall behind competitors who are using them.
Risk Management is Never Having to Say, “I’m sorry.”
Another reason why executives need to understand how AI technologies are used in business is to manage the risk of financial or reputational harm. A common culprit is personally identifiable information that is inadvertently or purposefully fed into an AI tool, which then “owns” your company’s data. It is all too easy for people to sign up for ChatGPT, Claude, Grok or another AI service and feed company sensitive data as part of a chat prompt. ?
Executives should assess why, how and when data can be leaked into an AI tool so they can lead the development of “rules of engagement” for employees and vendors to follow when they use AI tools to ensure that sensitive customer and corporate information isn’t leaked into the tool. Leaders in knowledge intensive industries also must manage potential negative customer responses when they find out that the work products they receive were produced by AI instead of human beings.
Customer concerns about AI-generated work products are exacerbated by the phenomenon of “AI hallucination,” where a large language model presents false or misleading information as fact. An example of this type of error that made national U.S. news was attorney Michael Cohen’s use of Google Bard to support termination of his supervised release from prison, generating fictitious legal case citations that his lawyer included in a legal briefing to a U.S. court.? As was the case in Cohen’s legal briefing, end users of AI can expect to be held accountable for inaccuracies contained in AI-generated work products.
Organizations that adopt large language model AI technology must invest in robust governance and controls to mitigate the risk of misinformation that is a function of the underlying architecture of a large language model. The key flaw is an LLM's ability to generate grammatically and syntactically correct text that has no “objective other than consistency with a prompt,” as noted by Yann LeCun, Chief AI Scientist at Meta.
Customers are “Going bye bye”
Finally, for organizations that are in expertise intensive businesses such as business consulting or industry analysis, AI tools can significantly reduce the need for these services. For example, why spend the money on an annual subscription to an industry analyst service when you can ask ChatGPT, “what is the latest hype cycle for AI?” and receive a summary of key AI technologies at each stage of the Gartner Hype Cycle? On the other hand, companies must navigate the ethics issues and liabilities associated with using proprietary information obtained from a free chat tool instead of the company that owns the trademark and copyright for the information.
As a consumer of end user AI tools, liability for the use of copyrighted content accessed from an AI tool has yet to be litigated in the courts. If our experience with litigation between search engine providers and establishment media content creators is predictive of future litigation over intellectual property ownership for content ingested into a large language model, this issue may take more than a decade to resolve in the U.S. court system.? While litigation works its way through the courts, providers of knowledge products must wrestle with risk to revenue and margin as potential customers access their content through AI tools rather than directly from the knowledge provider.
Conclusion: Balance Opportunity with Risk
In summary, today’s AI tools are easily and inexpensively accessible to customers, employees, and competitors. Ease of access to end user AI tools can create significant positive or negative outcomes for a company. Therefore, executives must understand AI use by each of these categories of stakeholders to maximize AI benefits while simultaneously mitigating its risks. They must also develop sufficient understanding of architecture and design of AI technologies to make wise decisions about where and when to incorporate AI into their business models.
LEADER 2: AI Winter is Coming (Kinda)
by Chris Morales - Business Consulting and Digital Content Development for Small Businesses, Young Professionals, & Creatives
If the hype of generative AI is falling, then the hype cycle over the forthcoming AI winter is in full force. But even if winter is coming, it won’t be a long one, and for good reason.
Without traveling down the road of previous hype cycles—Y2K, Web 1.0, or the many, many, many AI hype cycles of the past—I think it best to jump right into “why this is different.” (I would suggest a session with your favorite generative AI for details about past hype cycles.)“
...the spirit indeed is willing, but the flesh is weak. (Matthew 26:41)
If the winters of AI past were the by-product of weakness, that weakness was present not only in the technology but also in the consumer market mindset. While an inconceivable number of hours has been spent developing generative AI, those hours and “that community” have mostly been a closed circle with little to no appeal to the broader consumer market. Not until recently has AI’s role in our everyday lives been one of limitations and incremental improvements to “our convenience.”? But AI has been chiefly boring until now. People understood that it existed in things like ChatBots and worked under the alias of “Al Gorithm,” but they didn’t care how it worked. It did what the companies that controlled it made it do.
After Open AI released ChatGPT, handing the keys for a test drive to the broader market, something changed. But that narrative is nearly two years old, and the popular trend is that the boiling interest is beginning to simmer. AIs have seemingly found their niche in the hands of those of us who are determined to make them our co-intelligence and push them in ways their developers didn’t imagine (ethically, of course). The familiar belief is that funding rounds will see diminished returns, and the technology will fade into the background of the general consumer’s attention span.
However, this overstated belief fails to recognize that once, the consumer mindset needed to be stronger in its understanding of the impact of technology, but now this is not the case. The consumer market makeup now includes more of a generation that has grown up with product launches, annual releases, and regular software updates, which grasp the concept of iterative development and improvement. The potential and promises of AI are real for them, and they understand that, as Ethan Mollick says, “Today is the worst version of AI that you will ever use.”
The hype of AI may die down in business circles as more organizations slow down adoption without case studies of data quantifying ROI. Still, for the consumer market, it has been primed. The launch of Apple Intelligence (Novemberish?) will only further fold the consumer mindset into the methodologies and understandings of how to train AI in service of the broader market. So, if “Winter is Coming,” it will be brief because while the consumer market may not be ready for the broader impact of AI on society, it has an adoption mindset. Generative AI is as much a technology looking for a problem as it is a technology looking for a mindset, a mindset now more broadly available than ever before.
LEADER 3: Nope, NLPs Have Been Here for Decades
by Manoj Vadakkan - Management Consultant | Scrum Master| Artificial Intelligence | Generative AI | Creator of AI Scrum Consultant Stevieai.com/ | Conference Speaker
"The illiterate of the 21st century will not be those who cannot read and write, but those who cannot learn, unlearn, and relearn" Alvin Toffler (Future?Shock)
Is AI overhyped? Possibly to some extent, but the truth is that it has been integrated into our lives for a while. Machine learning and natural language processing (NLP) have quietly shaped our experiences through technologies like Amazon's recommendation systems and Netflix's personalized content suggestions. Having worked on NLP technologies myself in the '90s, I've seen firsthand how these early systems laid the groundwork for today's more advanced applications. Over time, customer service chatbots, virtual assistants like Alexa, and Apple’s Siri, which had limitations in their early versions, have improved significantly. With the release of ChatGPT in November 2022, AI became even more accessible to everyday users, bringing sophisticated language generation into mainstream use. While AI might sometimes be perceived as overhyped, its growing presence and evolving impact on our lives are undeniable.
AI has undoubtedly brought convenience and innovation, but it also poses significant challenges to the current job market. Many jobs that involve routine and repetitive tasks are being automated, leading to job losses in sectors such as manufacturing, data entry, and customer service. Customer service representatives are facing competition from more sophisticated chatbots that are getting better at handling basic queries, though they still struggle with complex interactions.
AI is increasingly automating tasks across knowledge worker professions too. In product management, AI is becoming increasingly valuable, not just in analyzing customer responses, complaints, and market trends to identify opportunities for new features or products, but also in the creative process of new product development. By helping to generate innovative ideas and formulating hypotheses for testing, AI enhances decision-making with greater rigor and creativity. Simultaneously, in programming, AI is evolving from simple code assistance and suggestions to more advanced roles, acting as an AI pair programmer that collaborates with developers, potentially transforming the way software is created and optimized. These examples are not isolated; they represent a broader trend that’s impacting a wide range of professions. As AI continues to evolve, it’s clear that the need to adapt extends to other sectors as well, such as healthcare, finance, and education.
How will these advancements affect the job market in these fields? Will AI complement the roles of product managers and programmers, or will it fundamentally alter the demand for these professions?
The concerns about AI’s impact on the job market echo the fears that arose during the computerization movement in India in the 1980s. In fact, 1984 was observed by some as the "Anti-Computerization Year," with widespread protests against the introduction of computers in industries like railways and banking. Many feared mass job losses, and I personally joined these protests, driven by the anxiety of potentially facing unemployment myself. Ironically, not long after, I enrolled in a master’s program in computer science, embracing the very technology I had once resisted. Despite the initial resistance, computerization moved forward, leading to workforce reskilling and new opportunities. This shift not only transformed these sectors but also helped India emerge as a global leader in IT. My experience shows that while new technologies can be disruptive, they often pave the way for growth and adaptation.
领英推荐
As AI and automation reshape the job market, the current workforce is understandably concerned, especially given the uncertain economy and shifting job landscape. However, many may not know how to respond to these changes, and even companies, despite outward appearances, are still grappling with their AI strategies. A solid starting point for both individuals and corporations could be basic education about AI, helping everyone understand its implications and potential. Meanwhile, students might not yet grasp the full extent of how AI will impact their future careers. Traditional roles like financial analysts, radiologists, and even physicians may not be in as high demand as they once were. Schools and universities must adapt quickly to prepare students for a future dominated by technology.
This uncertainty signals what Andrew S. Grove, in his book Only the Paranoid Survive, calls a "Strategic Inflection Point"—a critical juncture where the fundamentals of various professions are changing. Reflecting on my own experience, as I entered the job market when computer jobs were on the rise, I found inspiration in this book. It pushed me to learn as much as I could to stay relevant. Perhaps it's time to revisit that mindset, embracing the need to constantly learn and adapt in this new AI-driven era.
As we navigate the rapid advancements in AI and automation, the speed of change reflects the concerns Alvin Toffler raised in his book Future Shock. Toffler warned that the pace of technological change could overwhelm societies, leading to a sense of disorientation. If Toffler thought the changes of his time were significant, the exponential increase in technological advancements today makes our situation even more challenging. He emphasized the need for individuals and institutions to continually learn, unlearn, and relearn to adapt to these changes. By heeding these insights, we can turn potential confusion into opportunities for growth and resilience.
Are we prepared to embrace this change?
LEADER 4: Beyond the Hype Cycle
by Fred Miskawi - Vice President Expert - CGI Global AI Center of Expertise
When we think of hype in technology, we often picture a short-lived wave of excitement that quickly subsides. However, the AI revolution is different. It's not just a passing trend but a significant shift beginning to reshape many aspects of our society. While its true impact isn't always immediately visible, its effects are starting to emerge, fundamentally changing how we work, live, and interact. This transformation is in its early stages and represents the beginning of a silent transformation of IT and our interaction with computing.
Making an Exception for this Hype Cycle
The Gartner Hype Cycle, long considered the standard model for technological adoption, struggles to capture AI's unique progression. Technologies like AI, which have long maturity cycles and multiple rounds of innovations, require a different approach. Key milestones, such as ChatGPT's launch in November 2022, served as the hype cycle trigger. The early 2023 AI frenzy represented the peak of inflated expectations. Currently, we are experiencing a trough of disillusionment because expectations of the technology have outpaced the outcomes achieved in the market today. However, AI's slope of enlightenment seems to be occurring simultaneously with continued innovation, and the plateau of productivity is still to be defined.
AI's rapid advancement and broad applicability have disrupted this traditional model, suggesting that we should consider an exception to the Gartner Hype Cycle to better understand the impact of the technology, its growth path, and its future impact on our day-to-day lives.
Factors Driving AI's Unique Trajectory
Rapid Technological Advancement
In just 18 months, we've witnessed remarkable progress across multiple domains. Text generation has evolved from simple chatbots to complex essay and code writing. Image creation has advanced from basic manipulations to photorealistic syntheses, while video synthesis has progressed from crude animations to cinematic sequences. Additionally, 3D modeling for gaming and VR has become increasingly sophisticated.
The pace of AI advancement is unprecedented. For instance, in 2022, DeepMind's AlphaFold predicted structures for nearly all cataloged proteins known to science, totaling 200 million proteins. This breakthrough is accelerating drug discovery and our understanding of diseases. OpenAI's GPT-3, released in 2020 with 175 billion parameters, was followed by GPT-4 just three years later, rumored to have over 1.8 trillion parameters, demonstrating exponential growth in model complexity and capability. Moreover, the development of agent architectures and new cognitive algorithms will continue to accelerate the rate of AI advancement, pushing the boundaries of what these systems can achieve and how quickly they can evolve.
Transforming Software Development
Part of the reason why we will not see a lengthy trough of disillusionment before the next hype curve is the silent transformation happening in software development. AI is revolutionizing the software development lifecycle. Tools like GitHub Copilot are assisting in code generation, while AI-powered testing and debugging are accelerating quality assurance. Automated documentation is streamlining knowledge transfer. These advancements are estimated to boost developer productivity by 20-30% with the current version of the tooling. With upcoming models that will include stronger chain-of-thought and reasoning capabilities, that productivity boost will continue to grow rapidly. GitHub reported that developers who use GitHub Copilot complete tasks 55% faster than those who don't, with 60–75% of users reporting they feel more fulfilled with their job.
Societal Integration
The acceleration of coding means acceleration of outcomes for society and acceleration of the integration of the value provided by AI into society. AI builds upon and accelerates previous technological revolutions, including personal computers, internet connectivity, mobile devices, and cloud computing. The cumulative effect is a subtle but profound transformation of daily life and work processes. AI's integration into society is profound and accelerating.
In healthcare, Woebot, an AI-powered chatbot for mental health support, exchanges 4.7 million messages weekly with users from over 100 countries, providing accessible mental health support at scale. In education, CENTURY, an AI-powered learning platform, reported that students using their system for more than two hours across the year had an average grade improvement of three times the national average.
The AI Production-Absorption Paradox
A unique challenge emerges as AI's capability to innovate outpaces our ability to fully implement and adapt to these innovations. This creates a scenario where the technology's potential often exceeds our immediate capacity to leverage it fully. This is why, to keep up and avoid getting left behind, we need to focus on improving outcomes—delivering better quality with this new tooling sooner.
Focusing on Value and Societal Outcomes
Moving beyond conceptual discussions, it's crucial to shift our focus from merely making AI more powerful to making it more useful. AI should be seen not as an end goal but as a means to an end—a powerful tool, but a tool nonetheless. The economic impact of AI is becoming increasingly tangible, with PwC estimating that AI could contribute up to $15.7 trillion to the global economy by 2030. However, the real measure of AI's success will be in its ability to deliver better societal outcomes.
My colleague Diane Gutiw articulates this point well in her insightful article, "Let's stop talking about AI", where she argues for a shift in focus from AI as a concept to its tangible benefits and outcomes. She emphasizes the importance of leveraging AI to achieve meaningful societal benefits rather than just pushing the boundaries of its power.
Key areas where AI is demonstrating clear value include healthcare, where AI-assisted diagnostics are improving accuracy and speed; agriculture, where smart farming techniques are optimizing crop yields; and education, where personalized learning experiences are enhancing student outcomes. By concentrating on practical applications like these and measurable outcomes, we can move past the hype and engage with AI's real-world impact.
The Path Forward
The AI revolution is reshaping our world in significant ways that aren't always immediately visible but are profoundly transformative. While media attention may ebb and flow, the underlying advancements and their impact on society continue to surge forward. Our challenge isn't to manage inflated expectations but to adapt to a reality where continuous, transformative innovation is the new norm.
As we navigate this new landscape, our focus must shift from debating whether AI is overhyped to understanding and leveraging its evolving capabilities. The true measure of AI's impact will be in how it reshapes our work, our societies, and our approach to problem-solving in the years to come. We must be ready to rebuild our processes, skills, and even our understanding of what's possible in this new AI-driven world. This transformation is just beginning, and its full impact on IT and our interaction with computing is yet to unfold.
LEADER 5: Cuter Kittens Is All We Have To Show For the AI Hype?
by Charlie Fleet - SaaS Executive | M&A, Strategy & Product Ops | Authentic At Scale Team Builder
I’m starting to wonder if this AI thing is just a flash in the pan. Big tech has already spent $60,000,000,000 USD on AI in 2024 and all I get is cuter kittens??? I still scroll endlessly for something good to watch on Netflix and my dishes still need washing each night.
It's everywhere, I just had to look
Then I started to look around me… In 10 seconds zoom caught me up on the 30 minutes I missed… I seamlessly navigated my lack of Italian language skills on a family trip this summer with Google Translate in my back pocket… Instagram keeps my sons’ thumbs quite active as they scroll instantly.? It turns out that AI isn’t a monolithic thing in and of itself, instead it’s a set of technologies that for decades have been deeply embedded in our lives but is now seeing a massive surge in investment.
It's Actually Overwhelming
Here’s my problem: It’s all around me yet I have no control over:
I mean OpenAI didn’t ask me what features I wanted in Chat GBT 4o, did they ask you? How can I as a run-of-the-mill citizen who spends $4 a month on my phone backups compete against the tidal waves of investment from Google, Microsoft and the like?
As it turns out though I am in control. We are still very early days with AI capabilities and the companies are all experimenting with different technologies, partnerships, operating rules and ethical standards. ?The more I interact with AI today the more informed I am in questions around:
Hyping Up What I Want
With every day that I learn about AI the more able I am at being an informed citizen. I can make better choices around which AI enabled solutions I want to use and around how my company uses AI.
With any disruptive technology, it will eventually hit a hype cycle and with AI we are careening down to the through of disillusionment. It is during this phase that we call make the biggest difference.
With our informed choices and selective participation we can shape the role AI plays in our lives and our society at large.
Founder & CEO, Journey To Agility | Breaking down Al strategies, trends and tools for all tech professionals and enterprises
6 个月Great article with lots to chew on. "The illiterate of the 21st century will not be those who cannot read and write, but those who cannot learn, unlearn, and relearn" Alvin Toffler (Future Shock). What a powerful quote. Gonna steal this one :-)