It’s hard to escape news about Artificial Intelligence these days. From new discoveries in science to gaming to enhanced experiences on our phones, cars, etc, AI is constantly evolving. Excitement about AI is everywhere. It gets difficult to understand whether or not the hype is real. Gartner’s Hype Cycle provides clarity by highlighting major innovations in AI.
In this article, we look at the leading Artificial Intelligence innovations today according to Gartner’s Hype Cycle.
Gartner’s Hype Cycle
Gartner’s Hype Cycle is one of, if not the most, illustrious and highly relevant consultant models used by organizations all over the world to establish and constantly improve their technology strategies. It also influences the investment decisions of large enterprises. The Hype Cycle is essentially a graphical depiction of the maturity lifecycle of any new technology or innovation experience. It provides an objective map that covers development to adoption to decline and outlines the risks and opportunities of new technology.
The path of the cycle can be divided into five phases:
- Innovation/technology trigger
This is the public announcement stage where the new technology is introduced to the public via media events. It creates a hype about and garners industry interest in the new technology innovation. - Peak of inflated expectations
The next stage is characterized by excitement, enthusiasm, and a wave of buzz boosted by media coverage. There is a presumption for the new technology to go beyond the current reality of its potential. This leads to organizations investing in it without any solid strategy. Ergo, an investment bubble is formed. - Trough of disillusionment
Reality sets in and the initial over-enthusiasm disappears. This stage is marked by impatience for results, performance issues, slow adoption, and a failure to meet revenue expectations. Positive hype dies down as negative hype is born. - Slope of enlightenment
This stage is characterized by realistic expectations. Early adopters begin to see benefits and regain motivation to move forward. Contextual understanding of the new technology grows, thus outlining the product’s actual potential, limitations, and how it can deliver value. - Plateau of productivity
The final stage is the real-world adoption of the technology following a successful demonstration of its benefits and efficiency. There are high levels of productivity and innovation becomes mainstream.
The duration between the peak of inflated expectations and the plateau of productivity is known as the time-to-value gap.
Where is AI today?
The AI Hype Cycle is full of innovations anticipated to drive high benefits with many of them expected to hit commercial adoption in two to five years. AI innovations fall under four categories thus far.
1. Data-centric AI
Data-centric AI focuses on enhancing the data used to train the algorithm rather than the algorithm itself. It facilitates improvements in deploying AI and deep learning-based solutions in computer vision scenarios. Data-centric AI improves overall yield and accuracy.
Among other things, innovation in data-centric AI includes synthetic data – a class of data that is artificially generated. Synthetic data is highly significant as it does not require any personally identifiable information when training machine learning models. It also reduces cost and improves ML performance.
Knowledge graphs, data labeling, and annotation are the other innovations in this category. The former is expected to experience mainstream adoption in 5+ years and the latter in two to five years.
2. Model-centric AI
AI models are still just as important to make sure the output continues to enable us to take finer actions. A few of the innovations in model-centric AI include composite AI, generative AI, casual AI, physics-informed AI, deep learning, and foundation models. Casual AI and foundation models should see mainstream adoption in 5+ years whereas the others will see it in two to five years.
Composite AI is the amalgamation of various AI techniques to make learning more efficient. It allows the opportunity to solve a broader range of business problems across industries. Composite AI helps expand and improve the quality of AI applications.
3. Application-centric AI
AI cloud services, autonomous vehicles, decision intelligence, edge AI, smart robots, computer vision, and intelligent applications are some of the innovations present in this category. Computer vision is expected to reach mainstream adoption in less than two years while decision Intelligence, edge AI, AI cloud services, and intelligent applications are expected to do the same in two to five years. Smart robots should see mainstream adoption in five to ten years while autonomous vehicles should experience the same in more than ten years.
Edge AI is essentially the deployment of AI applications in devices throughout the physical world. It is a paradigm that brings computation and data storage close to the location of the device. Edge AI improves operational efficiency and enhances customer experience.
4. Human-centric
Innovations included here are digital ethics, AI trust, risk and security management (TRiSM), responsible AI, AI maker, and teaching kits. Responsible AI is expected to reach mainstream adoption in 5+ years while the others are expected to do the same in two to five years.
Responsible AI facilitates the right outcomes by solving problems based on delivering value versus tolerating risks. It is a framework for developing and deploying AI with good intentions. Responsible AI is rooted in scaling AI with transparency, safety, and fairness.
Conclusion
When it comes to AI, there is a multitude of technologies to consider. You need to explore the possibilities of each and figure out how it adds value to your organization and your customer. Additionally, keep a close eye on innovations expected to hit mainstream adoption in two to five years.
