Utilizing smaller, yet more distinctive datasets for each deployment could enhance effectiveness. We anticipate that more startups will demonstrate that reliance on major tech companies is unnecessary in 2022.

Human-powered AI is on the rise. Explainable AI (XAI) systems are expected to yield outcomes that are comprehensible to humans, with AI decisions becoming even more valuable due to human involvement. Varun Ganapathi, PhD, Co-Founder and CTO at AKASA, highlights the necessity of human presence for AI to be truly effective: “For AI to genuinely provide value and efficiency, human guidance is essential to push projects towards completion. Without oversight, AI struggles to reach its full potential. This trend is likely to grow. Ultimately, machines will report to humans, who will serve as the managers overseeing both human staff and AI systems that need to be trained for their specific tasks. Similar to humans, AI requires continuous learning to enhance performance.” Increased human engagement also fosters greater trust in AI, countering fears that AI may replace jobs or make decisions devoid of crucial human traits such as empathy and kindness. Expect that human involvement will lead to more valuable AI-driven decisions in 2022.

Avoiding vendor lock-in is a priority for the telecommunications sector, which is pursuing an innovation known as Open RAN. This initiative aims to free operators from dependence on specific vendors and encourage smaller competitors to challenge the current market dominance held by a few firms. Businesses are keen to prevent being ensnared by any AI vendor. Doug Gilbert, CIO and Chief Digital Officer at Sutherland, explains: “Early adopters of basic enterprise AI integrated within ERP/CRM platforms are starting to feel confined. In 2022, we will witness organizations taking proactive measures to avoid AI lock-in, and rightly so, as AI is inherently complex. When embedded in systems like ERP, organizations surrender control, transparency, and innovation to the vendor rather than maintaining it themselves. AI should not be seen merely as a product or feature; it is an evolving set of capabilities. To extract the most impactful results from AI, more enterprises will pivot towards a strategy that combines various AI capabilities to address specific challenges or achieve outcomes. This shift implies a growing demand for advanced and customizable AI solutions, with businesses likely to deprioritize or eliminate costly foundational AI features within their enterprise platforms altogether.” In 2022 and beyond, a trend towards AI solutions that mitigate vendor lock-in is anticipated.

Chatbots are also enhancing their capabilities. Many of us have felt the frustration of wanting to talk to a human instead of navigating a chatbot. “Today’s chatbots, while beneficial, have considerable limitations. Natural language processing is set to be surpassed by neural voice technologies that enable near real-time natural language understanding (NLU),” comments Gilbert. “With capabilities to grasp complex sentence structures and even emotional nuances, NLU will refine the accuracy and overall experience of conversational AI.” This advancement may lead to two significant outcomes: real-time augmentation of human assistance by suggesting responses based on user behavior or expertise, and a more positive customer experience facilitated by a more natural interaction through NLU. In 2022, chatbots are poised to offer a more human-like interaction.

It’s not merely about quantity, but the quality of data in AI systems. A successful AI framework relies on a functioning model and high-quality training data. Collecting vast amounts of data is futile if it’s not accurately labeled and of high caliber. Gabriel Straub, Chief Data Scientist at Ocado Technology, states: “Andrew Ng has emphasized data-centric AI, highlighting that enhancing data quality can yield better results than focusing solely on algorithm improvements, at least for equivalent effort levels. How can this be practically implemented? Ensuring that data consumers remain central to your data strategy and that data governance acts as a facilitator for safe value extraction from data, rather than locking down data, will significantly enhance outcomes.” Expect that in 2022, the AI sector will prioritize data quality.

Want to delve deeper into AI and big data insights from industry experts? Check out the AI & Big Data Expo. Upcoming events will take place in Santa Clara on May 11-12, Amsterdam on September 20-21, and London on December 1-2.