Feature
A lot of the commentary about generative AI has so far focused on what it can do for business and the productivity gains it can bring to all areas of the digital workplace. What remains to be seen, however, is when all this is going to happen.
In its August 2023 Hype Cycle for Emerging Technologies, Gartner forecasted it could take two to five years before generative AI could offer organizations the truly transformational benefit they expect from the technology.
The report, which takes key insights from more than 2,000 technologies and applied frameworks that Gartner profiles each year, includes some technologies that are even further along their development cycle, with some expected to take as long as 10 years to gain traction in organizations. But Gartner's time horizon prediction about generative AI stands out because of the massive investment and engagement in the technology we are seeing now.
So, is it too early to invest in generative AI? Experts weighed in.
Don't Ignore the Hype
John Sviokla, co-founder of GAI Insights, a research organization focused on the strategic adoption of GenAl, said the problem with generative AI predictions right now is that they are overly general. In his view, the technology is already delivering on its productivity promise, just not everywhere you'd expect it to deliver.
To arrive at this conclusion, he — along with GAI Insights co-founder Jimmy Hexter and GAI Insights CEO and co-founder Paul Baier — defined a concept they call WINS, which they presented in an article published in the Harvard Business Review:
"Our case studies, based on our growing global community of over 3,000 GenAI practitioners, point to a new category of work, more precise and actionable than 'knowledge work.' We call it WINS Work: the places where tasks, functions, possibly your entire company or industry are dependent on the manipulation and interpretation of Words, Images, Numbers, and Sounds (WINS). Heart surgeons and chefs are knowledge workers but not WINS workers. Software programmers, accountants, and marketing professionals are WINS workers."
In an interview with Reworked, Sviokla said that to view generative AI in the context of the hype cycle creates an excuse for management to do nothing, which carries significant risk, particularly for organizations that utilize a lot of WINS work.
“If you’re running a law firm, a marketing firm, an accounting firm, a tech-enabled services firm, and you are not both enthused and scared to death about what [generative AI] can do today, never mind what it can do in 12 months, you are putting your head in the sand,” he said. “If anything, these firms need to pay even more attention than they are already.”
When it comes to the investment and adoption of generative AI, Sviokla said organizations should think of the risk/return in terms of a continuum of value — or a gradual way of getting to the maximum value of the technology.
For instance, organizations can get started using publicly available tools before introducing gatekeeping/security software to manage the interactions with the LLMs, and, down the road, build their own content using an open source model. Doing so, he said, means companies can get very quick returns for early endeavors.
Sviokla believes that in the medium term — two to five years — there will be new "verticalization" because generative AI makes it easier to light up the unstructured or semi-structured data that organizations have. It also makes custom integration software easier and cheaper to write.
“I’d expect that there will be more competitive advantage created by people who can create a technology/learning system to unleash proprietary knowledge for value,” he said.
Related Article: Generative AI, the Great Productivity Booster?
Be Patient, Breakthroughs Are Inevitable
While Gartner's five years is realistic, it is also a conservative assessment, said Alok Shankar, engineering manager for Oracle Health AI. Generative AI has been advancing rapidly, he said, but whether its full potential can be realized within five years depends on several factors.
Progress in AI is fast-paced, with notable achievements in content generation and creativity. However, complex and context-specific tasks may take longer to perfect.
“Challenges vary across domains, and ethical and regulatory considerations will play a role. While generative AI is progressing, its widespread adoption may extend beyond five years, especially for intricate tasks,” Shankar said. “Nonetheless, continuous innovation suggests significant gains in the foreseeable future.”
John Pennypacker, VP of sales and marketing at Deep Cognition, a firm that develops generative AI platforms and solutions, said it can be frustrating to wait for significant progress in generative AI, but the field is complex and requires substantial advancements. And the timeline for such advancements depends on several factors, including research breakthroughs, computational resources and the evolving nature of AI technologies.
As an example, he points to the rate of progress in areas like natural language processing (NLP) and computer vision. “In the last five years, we have seen tremendous advancements in NLP, with models like GPT-3 demonstrating remarkable improvements in language understanding and generation,” he said.
“Similarly, advancements in computer vision models like convolutional neural networks (CNNs) have led to substantial gains in image recognition and understanding."
Although waiting for advancements in generative AI might seem long, breakthroughs can happen unexpectedly, leading to exponential improvements. Patience, Pennypacker said, is crucial in AI, and even relatively short timeframes can yield remarkable advancements that reshape how we interact with technology and data.
Related Article: Generative AI Is Cool, But It Isn't Corporate (Yet)
Learning Opportunities
Webinar
May
16
Advance Interoperability in Healthcare: How to Navigate Barriers and Improve Clinical Quality Outcomes
Discover how enhancing your interoperability strategy can drive better outcomes for both patients and staff.
Register
Webinar
May
21
Navigating the GenAI Era: Skills to Scale and Strategies to Stay Ahead
Join us for an insightful conversation as we delve into the world of Generative AI and its profound impact on the workforce.
Register
4 Current Generative AI Productivity Gains
Iu Ayala, founder and CEO of AI consultancy Gradient Insight, said the timeline to productivity gains can vary depending on several factors such as industry adoption, regulatory considerations and the specific application of generative AI. Five years can be a realistic timeframe for certain applications, while others might witness earlier or later gains.
But to address the question of whether Gartner's five-year timeline for productivity gains is realistic, he said it's essential to consider both the current state of generative AI and the practical applications that are already demonstrating productivity improvements. He lists four such applications:
1. Content Generation and Automation
Generative AI models, such as GPT-3 and its successors, have shown remarkable progress in content generation, Ayala said. These models can assist in writing articles, reports and even code, thereby significantly reducing the time and effort required for these tasks. Content generation is a tangible example of productivity enhancement that's already in use across various industries.
2. Enhancing Creativity in Design
In sectors like graphic design and product prototyping, generative AI is being used to assist and augment human creativity. Tools like Adobe's Project Aero and NVIDIA's GauGAN enable designers to generate design concepts rapidly, streamlining the creative process.
3. Healthcare and Drug Discovery
Generative AI is making significant strides in healthcare, aiding in drug discovery by simulating molecular structures and predicting potential drug candidates, Ayala said. This accelerates the development of life-saving medications and demonstrates substantial productivity gains in a critical field.
4. Natural Language Processing in Customer Service
Conversational agents powered by generative AI are being employed for customer service interactions. These agents, Ayala said, can handle routine inquiries, freeing up human agents to focus on more complex and nuanced customer issues, thereby improving overall efficiency.
“The key takeaway is that generative AI is poised to revolutionize multiple industries, and businesses that strategically integrate these technologies can expect substantial gains in productivity,” Ayala said.
Related Article: Is It Too Early to Talk about the Power of Quantum Computing and Generative AI?
Software Development
One area has observed clear and immediate tangible productivity gains, according to Fabrice Bellingard, VP of product at code-building firm Sonar: software development.
While for now the tools are confined to addressing common, repetitively documented and low-complexity coding issues, he anticipates more advancements — and more quickly than five years out — that will make it possible to produce code faster, speeding up software production timelines.
But, he said, we shouldn't hurry the process along too rapidly.
“Yes, speed matters and AI is rapidly empowering developers, allowing them to elevate their focus, innovate and contribute more business value by automating routine tasks,” he said. “But we at Sonar urge organizations to consider the importance of software quality and take immediate action to make sure that the code comprising any software built by that organization is clean and issue-free."
McKinsey famously noted that every company today is a software company. This, Bellingard said, means that businesses are innovating and competing on software, with a development process that will reap the transformational benefits of AI in under five years. Yet, they cannot risk cracks in the software foundation that would turn their code into what he believes is the biggest business liability today.
“The same cautionary advice applies to any realm that AI touches,” he said. “We need humans to guide the effort and to make sure that we are not sacrificing quality for any benefits of AI."
Regarding the human element, some say AI is a threat to developer jobs. Bellingard disagrees. "In fact," he said, "I see indicators of transformational AI benefit when it comes to technical career advancement.”