2022 Technology Predictions for AI in the Enterprise
The global use and further development of AI continued to grow in 2021 as enterprises found more ways to deploy it and developers discovered new ways to capture its possibilities for business users.
So, what might 2022 bring for AI and a wide range of related IT fields from MLOps to security, cloud and edge computing, open source, the metaverse and more?
To answer that question, we received a wide range of predictions from IT industry experts who shared their thoughts with EnterpriseAI. We are publishing them here, edited for clarity and brevity, to give our readers an early look at what may come in 2022 in enterprise AI and related technologies.
Rodrigo Liang, the CEO and co-founder of AI platform vendor SambaNova Systems, said he sees companies moving away from DIY AI and linking up with vendors who can help them better reach their business goals.
“Organizations recognize the talent shortage and internal inertia to technology advancements adds significant performance risk,” said Liang. “As a result, traction for AI-as-a-Service expands allowing for investment in applying insights over creating those insights. How many companies are going to be able to afford to hire hundreds of people just to manage one model and have thousands of GPUs interconnected to run one thing?”
In addition, business verticals including banking, financial services, insurance and manufacturing will further deploy AI in transformative ways to bolster their performance and operations and will never look back, said Liang. “Just like how the internet changed every facet of commercial transactions, AI will have the same level – if not more – of an impact. These industries will move from test deployments to production and will earn the returns promised with AI.”
Mark Brayan, the CEO of machine learning data training vendor Appen, said that in 2022 he sees the idea of “responsible AI” shifting from an aspiration to a foundational requirement for most AI projects.
“In 2021, responsible AI was one of the hottest topics in the AI industry, but adoption remained relatively low,” he said. “In 2022, however, the stakes become much higher, as businesses recognize that responsible AI leads to better business outcomes. The principles of responsible AI are now well-established: unbiased data, fair treatment on the data collection and labeling side of the industry, and a recognition that AI projects should promote the social good, or at least avoid the potential for social harm. Implementing these principles ensures that AI projects work as expected and protects the brand.”
In addition, said Wilson Pang, Appen’s CTO, the new year will also see model evaluation and tuning going mainstream for enterprises.
“In 2022, the need for regular model evaluation and tuning becomes AI program table stakes,” said Pang. “Machine learning models are dynamic – they cannot be deployed and forgotten. ML models in production need to be updated and retrained based on a variety of factors, including the ongoing results, as well as changes in infrastructure, data sources and business models.”
Bill Scudder, the AIoT (artificial intelligence of things) general manager at industrial optimization software vendor AspenTech, said he sees AI’s maturation into industrial AI reach full bloom in 2022, graduating to real-world product deployments with concrete time-to-value.
“To achieve this, we will see more industrial organizations make a conscious shift from investments in generic AI models to more fit-for-purpose, precise industrial AI applications that help them achieve their profitability and sustainability goals,” said Scudder. “This means moving away from AI models that are trained on large volumes of [facility] data that cannot cover the full range of potential operations, to more specific industrial AI models that leverage domain expertise for interpreting and predicting with deep analytics and machine learning. Industrial data will be transformed into successful business outcomes across the full asset lifecycle.”
Scudder said he also expects to see progress as executive buy-in and cultural change within organizations accelerate industrial AI deployments.
“Digital executives like chief digital officers (CDOs) will be crucial to overcoming these obstacles,” he said. “CDOs will have a unique role to play in shepherding digital transformation and industrial AI through their organization – bridging the gap between legacy systems and new technologies, fostering collaboration across silos, and shifting from mass data collection to strategic industrial data management. All these duties will be essential to ensure that an industrial organization can execute a digital transformation plan that sees wider adoption of, and strategy around, fit-for-purpose industrial AI applications.”
Omer Har, the co-founder and CTO of external data platform vendor Explorium, said he sees MLOps (machine learning ops) move from the periphery to the center of DevOps, data and ML practices.
“The pandemic gave many large organizations the push they needed to embrace AI and analytics,” said Har. “These predictive models have assumed critical importance in industries ranging from insurance to consumer packaged goods. The key to their predictive power is a constantly refreshed stream of external training data. That means the models must be frequently retrained and seamlessly redeployed to production. For many DevOps professionals, this process has become as mission-critical as traditional SaaS deployments and require the same level of instrumentation and careful tool selection. MLOPs is no longer a sideshow. In many organizations, it is quickly becoming the main show.”
Ali Siddiqui, the chief product officer for BMC Software, said he sees similar progress for AIOps (AI operations) inside enterprises this year.
“AIOps will grow in 2022 as businesses adapt to be successful in delivering the digital experiences customers demand as they move to hybrid cloud environments,” said Siddiqui. “AIOps will provide organizations with insights into their data to help them identify pain points, reduce noise, provide visibility to issues before they impact the business and meet business objectives while saving time and money. AIOps also eliminates the need to analyze thousands of events and transforms large amounts of data into actionable information which is key for business success and efficiency.”
Matt Watts the chief technology evangelist at data-centric software vendor NetApp, said he expects to see more progress in AI adoption at the edge in the new year.
"The adoption of AI technology at the edge of networks will continue to accelerate across the manufacturing, transportation, agriculture, entertainment and hospitality industries,” he said. “For example, the agriculture and food processing industries will use AI for harvesting and packaging. An explosion in tiny machine learning chipsets for low-cost and resource-constrained devices such as remote sensors that can collect and process data at the edge will drive this trend. These chipsets will fuel the ever-growing edge-core-cloud data pipeline, which industries will need to access and leverage to differentiate themselves in the marketplace.”
In addition, Watts said he expects to see more forward progress in the use of quantum computing in 2022. “Quantum computing will accelerate as major IT players and startups use the technology for complex tasks such as drug discovery, financial risk calculation, automotive battery design, and supply chain optimization,” he said. “Organizations will be more vocal about their quantum computing strategy in 2022, sharing how they will deliver quantum computing as a service to their customers and overcome challenges such as building a data pipeline into the quantum computing cloud.”
Peak co-founder and CEO Richard Potter said he sees the nascent field of decision intelligence, which he called the commercial application of AI to decision-making processes, as the most important B2B movement of a generation. “We are at the stage of ‘narrow AI,’ where machine learning and AI can make predictions and categorizations for specific purposes. But to solve the biggest business challenges, AI needs to be focused on an outcome, on delivering against business objectives and driving tangible results. Businesses that make great decisions consistently win. That is why decision intelligence is how most businesses will adopt AI.”
Alicia Frame, the director of product management for data science at graph database vendor Neo4j, said that in 2022 companies must embrace the role of the citizen data scientist for employees whose primary job functions lie outside the field of data science and analytics. The data science field is one of the fastest growing, and with the workforce currently experiencing ‘The Great Resignation,’ companies will need to make data science more accessible to help fill gaps on their teams,” said Frame. Over the past five years, the inquiry volume of clients striving to learn about digital ethics has more than tripled. Graphs are built for providing context to systems, which allows for increased explainability in AI/ML systems. As more organizations explore the different technologies to reach this point, 2022 will be a turning point for many organizations as they leverage graph technology to enhance their ability to address bias and create more ethical AI/ML systems.”
Steven Mih, the co-founder and CEO at Ahana, which provides managed Presto services on AWS to help simplify open data lakes analytics, said he expects to see investments and adoption of managed services for open source to soar in the new year. “More companies will adopt managed services for open source in 2022 as more cloud-native open source technologies become mainstream, including Spark, Kafka, Presto, Hudi and Superset,” he said. “Open source companies offering easier-to-use, managed service versions of installed software enable companies to take advantage of these powerful systems without the resource overhead so they can focus on business-driven innovation.”
Manjusha Madabushi, the chief technology officer at Talentica Software, said he sees big changes coming in the world of metaverses in 2022. "With Facebook renaming itself as Meta and starting to build the metaverse, we will see big investments and innovation in virtual world-based online gaming, social networking and virtual reality based products like virtual conference platforms. Expect new applications that enable meetings with characters in the virtual world filling the place of corresponding human entities.”
Disruptions will come in 2022 when it comes to cyberattacks on already-pressured companies including chipmakers, said James Carder, the chief security officer and vice president of labs at IT security vendor LogRhythm. “A leading country producing semiconductor chips will have its supply-chain compromised, resulting in major shortages of critical materials,” said Carder. “As we have seen with the pandemic, cybercriminals will take advantage of periods of societal disruption to manipulate companies and governments for financial gain. The global chip shortage, which shows no sign of slowing down … is another period of disruption that hackers will soon exploit. As countries seek to ramp up production, one country will be caught attempting to corner the market by using fraudulent methods to gain access to the production and supply of the leading chip-producing countries.”
Carder also predicts that the supply chain of a major vaccine manufacturer will be halted by ransomware. “In 2021, ransomware attacks crippled Colonial Pipeline and JBS. In 2022, cybercriminals will set their sights on conducting a ransomware attack against one of the pharmaceutical companies producing the COVID-19 vaccine. This will interrupt the production of critical booster shots and keep other lifesaving drugs from reaching patients. The resulting fallout will fan the flame for foreign and domestic vaccine disinformation campaigns.”
Murli Thirumale, the vice president and general manager of the cloud native business unit at flash storage vendor Pure Storage, said he sees 2022 as the year when containers will become a technology staple of mid-market companies and not just used by huge corporations with large IT staffs. “Previously, Global 2000 companies were those that could afford to experiment and deploy newer technologies like containers. After all, they have the means to staff talented DevOps teams and invest in multi-year transformation initiatives. However, as containers move past the stages of initial innovation and adoption, and the industry moves into early maturity, mid-market companies will begin deploying and experimenting with this technology even more. Not only will containers work better out of the box, they will be delivered as a service and consumed with ease.”