AI Predictions and Innovations for 2022
Most enterprises failed to scale their AI deployments this year. Will 2022 be the year when this actually happens?
- By Leah Forkosh Kolben
- November 30, 2021
Companies had high hopes in 2021 that artificial intelligence (AI) would enable them to make better and faster predictions to gain competitive advantage. However, Accenture research reported that despite their best efforts, nearly 80 percent of enterprises last year failed to scale AI deployments across the organization.
However, we see AI evolving quickly to begin delivering on its promise in 2022. We predict a new wave of technological advancements that will help companies overcome the common challenges for operationalizing AI.
Here are the innovations we expect to see in the year ahead.
Relief to Data Scientists Bogged Down by Tedious Tasks
This year, data scientists spent more than half their time cleaning data, integrating disparate storage systems, and finding the needed storage capacity and processing power to put AI and machine learning (ML) models into production. A great deal of highly trained human resources were spent solving operational issues because most models were built using a sterile environment with clean data. When models were put into production, they didn’t perform as intended due to a lack of awareness of the complexities of integrating real data.
In 2022, we expect some relief for data scientists as more and more machine learning tasks become automated. Many labor-intensive tasks (such as preparing data, engineering features, and training models) that involve repetitive, tedious, and time-consuming functions will begin to be automated in the coming year. Automating the process results will not only provide relief to data scientists by letting them focus on what they do best -- perfecting algorithms -- but will also enable IT teams to put higher performing AI/ML models into production, faster.
AI Will Become More Accessible to Developers
Previously, only big players such as Google and Facebook had the deep pockets to make AI/ML models a reality. In 2022 there will be more off-the-shelf technology used by developers at midsize companies to make AI/ML models more accessible. There will be functionality readily available to make applications talk, convert speech to text, automate video and image analysis, automatically eliminate inappropriate or illegal content, and many other industry-specific use cases.
In addition, there will be fewer complexities when choosing and porting AI software to different platforms, as bring your own compute and storage (BYOCS) will become a reality. With AI/ML models consuming ever more resources, AI developers will be able to choose the best-of-breed compute and cloud solution per machine learning model. Selecting the most efficient platform on the fly will prevent vendor lock-in and provide the flexibility data scientists need to optimize computing resources while giving them the luxury to try new, innovative technologies without risking their whole AI/ML environment.
AI Will Evolve More Rapidly
In 2021, most companies were still in the experimentation or proof-of-concept phase of AI. In the coming year, we are going to see a shift towards AI-first approaches. AI applications will be at the forefront of enterprise (and even government) strategies. As AI/ML models become the norm, companies will experience a steeper curve of improvements, expanding AI to become part of every department and impacting every business process.
Deloitte's State of AI report found that about 80 percent of business and IT executives believe AI will be critical to their company's success over the next two years. In addition, about three-quarters of the respondents believe all businesses will use AI in the next three years. We expect to see the wider adoption of AI technologies in several industries but primarily in healthcare, retail, manufacturing, and finance.
In 2022, the use of AI will not only be more prevalent, it will also be more strategic. AI will continue to be used to achieve productivity gains. In addition, in the coming year AI will also be used to rethink and redesign products, services, business models, and overall strategy. AI won’t be something tacked onto the existing infrastructure but instead will be an integral part of a company’s technology stack, providing insights for customers, partners, and employees in real time.
With the new online economy, companies will race to use AI insights to become more competitive by being truly data driven in 2022. The increase in research and development driven by the giant companies will result in more innovation and enablers bringing AI within reach for smaller companies as well.
This year’s challenges of integrating, cleaning, and processing data will continue, but at the same time there will be a flood of more generic open-source solutions that can replace manual tasks, freeing up data scientists for more strategic tasks. As the increasingly online economy demands more data for companies to be more efficient and provide a better customer experience, there will be more enabling technologies available to ease the transition to an AI economy.
Leah Forkosh Kolben is the co-founder and CTO of cnvrg.io. Kolben completed her BSc in mathematics and physics already in high school, worked as a software team leader at WatchDox (acquired by Blackberry) while earning a BSc in Computer Science, and founded cnvrg.io where she pursues technological innovation as well as gender equality for women in STEM positions. You can reach the author via email or on LinkedIn.