Gartner recently updated its Hype Cycle for Artificial Intelligence 2021 report by Gartner analysts Shubhangi Vashisth and Svetlana Sicular.
Several AI technologies like computer vision, NLP, chatbots, and edge AI have been driving enterprise adoption over the years. The Hype Cycle for Artificial Intelligence 2021 presents some new trends that will dominate the AI landscape. 34 AI technologies are described in this year’s report, with an above-average number of innovations reaching mainstream adoption within two to five years.
Most organizations have adopted proven AI technologies like NLP and Computer Vision to drive growth, improve customer experience and create new products and services. But their focus continues to be on improving speed to market, productionizing AI, and going beyond POCs. Hence, the following four megatrends dominate this year’s AI landscape:
Operationalizing AI initiatives
By 2025, 70% of businesses are expected to have operationalized AI architectures. Companies are looking to operationalize AI platforms to enable scalability and accelerate AI adoption and growth. AI orchestration and automation platforms and model operationalization reflect this trend. ModelOps reduces the integration cycle into production, optimizes operations, and achieves a higher degree of success.
Efficient use of data, models, and compute
Efficient use of data, models, compute power, and other resources lead to innovations in AI. Multi-experience AI, composite AI, generative AI, and transformers are examples of this trend. For instance, composite AI combines elements for deep learning, graph analysis, and other techniques to solve a broader range of business problems in a more efficient manner.
As with most technologies, there is the positive and negative impact of AI. While some results can be misleading and biased, others can lack transparency and auditability, seriously affecting an organization’s reputation and customer base. The AI landscape is moving towards developing a fair and responsible AI. This includes explainable AI, risk management, and AI ethics for increased trust, transparency, fairness, and auditability.
Data for AI
As user behavior and patterns change over time, AI and ML models must be continuously updated to avoid becoming obsolete. Changes in regulations or global events like the Covid-19 crisis change the conditions in which businesses are conducted and completely alter consumer behavior. Such changes have resulted in a shift from traditional big data to what is known as “small and wide data”. Gartner expects that by 2025, 70% of businesses will shift their focus to “small and wide data” which will enable more robust analytics and AI.
The report also identifies six key technologies to watch out for in the innovation trigger phase; and expected to hit the plateau of productivity (the end of the hype cycle) within two to five years:
Composite AI - This approach combines various AI techniques to solve business problems more efficiently. The right choice of AI techniques depends on the business problem and the data sets available. According to the report, composite AI has penetrated 5 to 20% of the target market.
AI orchestration and automation platform (AIOAP) - This technology unifies development, delivery, operation, and model performance. Companies are combining DevOps and CI/CD elements into the AI model lifecycle to standardize DataOps, MLOps, and deployment pipelines. AIOAP has reached 1% to 5% of the target audience.
AI governance - AI governance establishes accountability around the use of AI. Governments around the world are setting regulations for AI with some voluntary guidance and some binding. This technology has reached 1% to 5% of the target audience.
Generative AI - This technology applies what it has learned to generate new content. Gartner predicts that this technology can potentially disrupt software coding and automate up to 70% of the work done by programmers. It is still early days for this technology with a reach of less than 1% of the target audience.
Human-centered AI – Also known as augmented intelligence, machines, and humans work together for greater efficiency and reliability. This has reached 5% to 20% of the target audience.
Synthetic data - This is one solution to the challenge of obtaining real-world data to train AI models. While it is cheaper and faster to get, the data generated can have bias problems and miss natural anomalies. This technology is emerging and has reached 1% to 5% of the target audience.
Read the complete report by Gartner here
About the Author
Kingshuk Ghosh, Product Manager, Kanverse.ai