Key Artificial Intelligence (AI) Trends

Artificial intelligence (AI) has taken on many new shapes and use cases as experts learn more about what’s possible with big data and smart algorithms. 

Today’s AI market, then, consists of a mixture of tried-and-true smart technologies with new optimizations and advanced AI that is slowly transforming the way we do work and live daily life. 

Read on to learn about some artificial intelligence trends that are making experts most excited for the future of AI:

5 Trends to Watch in AI Solutions

1. Intelligent and hyper-automated business processes
2. Emphasis on responsible AI development
3. AI as a tool for global good
4. AI and IoT working together
5. The emergence of decision intelligence

More on the AI market: Artificial Intelligence Market

1. Intelligent and hyper-automated business processes

With its ability to follow basic tasks and routines based on smart programming and algorithms, artificial intelligence is becoming embedded in the way organizations automate their business processes. 

AIOps and MLops are common use cases for AI and automation, but the breadth and depth of what AI can automate in the enterprise is quickly growing.

Bali D.R., SVP at Infosys, a global digital services and consulting firm, believes that AI is moving toward a certain level of hyper-automation, partially in response to the unexpected changes in manual data and procedures caused by the pandemic.

“We are in the second inflection point for AI — as it graduates from consumer AI, towards enterprise-grade AI,” D.R. said. “Being exposed to an over-reliance on manual procedures, such as mass rescheduling in the airline industry, unprecedented loan applications in banks, etc., the industries are now turning to hyper-automation that combines robotic process automation with modern machine learning to ensure they can better handle surges in the future.”

Although AI automation is still mostly limited to interval and task-oriented automation that requires little imagination or guesswork on the part of the tool, some experts believe we are moving closer to more applications for intelligent automation.

David Tareen, director for artificial intelligence at SAS, a top analytics and AI software company, had this to say about the future of intelligent automation:

“Intelligent automation is an area I expect to grow,” Tareen said. “Just like we automated manufacturing work, we will use AI heavily to automate knowledge work. 

“The complexity comes in because knowledge work has a high degree of variability. For example, an organization will receive feedback on their products or services in different ways and often in different languages as well. AI will need to ingest, understand, and modify processes in real-time before we can automate knowledge work at large.”

AI, automation, and the job market: Artificial Intelligence and Automation

2. Emphasis on responsible AI development

Because of the depth of big data and AI’s reliance on it, there’s always the possibility that unethical or ill-prepared data will make it into an AI training data set or model. 

As more companies recognize the importance of creating AI that conducts its operations in a compliant and ethical manner, a number of AI developers and service providers are starting to offer responsible AI solutions to their customers.

Read Maloney, SVP of marketing at H2O.ai, a top AI and hybrid cloud company, explained what exactly responsible AI is and some of the different initiatives that companies are undertaking to improve their AI ethics. 

“AI creates incredible new opportunities to improve the lives of people around the world,” Maloney said. “We take the responsibility to mitigate risks as core to our work, so building fairness, interpretability, security, and privacy into our AI solutions is key.” 

Maloney said the market is seeing an “increased adoption of the core pillars of responsible AI,” which he shared with Datamation:

• Explainable AI and interpretable ML: The ability to explain a model after it has been developed and providing transparent model architectures, which allows human users to both understand the data and trust results.
• Ethical AI: Provides sociological fairness in machine learning predictions (i.e., whether one category of person is being weighted unequally and eliminating historical human bias).
• Secure AI: Debugging and deploying ML models to keep security and privacy at the forefront.
• Human-centered AI: Where AI learns from human input and collaboration. Systems are continuously improving because of human input and bridging the gap between human and machine.
• Compliance: Ensuring AI systems meet the relevant regulatory requirements or regulations.

Companies are exploring several ways to make their AI more responsible, and most are starting with cleaning and assessing both data sets and existing AI models.

Brian Gilmore, director of IoT product management at InfluxData, a database solutions company, believes that one of the top options for model and data set management is distributed ledger technology (DLT).

“As attention builds around the ethical and cultural impact of AI, some organizations are beginning to invest in ancillary but important technologies that utilize consensus and other trust-ensuring systems as a part of the AI framework,” Gilmore said. “For example, distributed ledger technology provides a sidecar platform for auditable proof of integrity for models and training data. 

“The decentralized ownership, distribution of access, and shared accountability of DLT can bring significant transparency to AI development and application across the board. The dilemma is whether for-profit corporations are willing to participate in a community model, trading transparency for consumer trust in something as mission critical as AI.”

See more: The Ethics of Artificial Intelligence (AI)

3. AI as a tool for global good

Up to this point, AI has most frequently been used to optimize business processes and automate some home routines for consumers. 

However, some experts are beginning to realize the potential that AI-powered models can have for solving global issues.

Read Maloney at H2O.ai has worked with people from a variety of industries to brainstorm how AI can be used for the greater good.

“We work with many like-minded customers, partners, and organizations tackling issues from education, conservation, health care, and more,” Maloney said. “AI for good is fundamental to not only our work, including current work on climate change, wildfires, and hurricane predictions, but we are seeing more and more AI for good work to make the world a better place across the AI industry.”

Some of the most exciting applications of altruistic AI are being implemented in early education right now.

For instance, Helen Thomas, CEO of DMAI, an AI-powered health care and education company, offers an AI-powered product to ensure that preschool-aged children are getting the education they need, despite potential pandemic setbacks:

“On top of pre-existing barriers to preschool education, including cost and access, recent research findings suggest children born during the COVID-19 pandemic display lower IQ scores than those born before January 2020, which means toddlers are less prepared for school than ever before. 

“DMAI DBA  Animal Island Learning Adventure (AILA) is changing this with AI. [Our product] harnesses cognitive AI to deliver appropriate lessons in a consistent and repetitious format, supportive of natural learning patterns

“Recognizing learning patterns that parents might miss, the AI creates an adaptive learning journey and doesn’t allow the child to move forward until they’ve mastered the skills and concepts presented. This intentional delivery also increases attention span over time, ensuring children step into the classroom with the social-emotional intelligence to succeed.”

More on this topic: How AI is Being Used in Education

4. AI and IoT working together

Internet of Things (IoT) devices have become incredibly widespread among both enterprise and personal users, but what many tech companies still struggle with is how to gather actionable insights from the constant inflow of data from these devices. 

AIoT, or the idea of combining artificial intelligence with IoT products, is one field that is starting to address these pools of unused data, giving AI the power to translate that data quickly and intelligently.

Bill Scudder, SVP and AIoT general manager at AspenTech, an industrial AI solutions company, believes that AIoT is one of the most crucial fields for enabling more intelligent, real-time business decisions.

“Forrester has noted that up to 73% of all data collected within the enterprise goes unused, which highlights a critical challenge with IoT,” Scudder said. “As the volume of connected devices — for example, in industrial IoT settings — continues to increase, so does the volume of data collected from these devices. 

“This has resulted in a trend seen across many industries: the need to marry AI and IoT. And here’s why: where IoT allows connected devices to create and transmit data from various sources, AI can take that data one step further, translating data into actionable insights to fuel faster, more intelligent business decisions. This is giving way to the rising trend of artificial intelligence of things or AIoT.”

5. The emergence of decision intelligence

Decision intelligence (DI) is one of the newest artificial intelligence concepts that takes many current business optimizations a step farther, by using AI models to analyze wide-ranging sets of commercial data. These analyses are used to predict future outcomes for everything from products to customers to supply chains.

Sorcha Gilroy, data science team lead at Peak, a commercial AI solutions provider, explained that although decision intelligence is a fairly new concept, it’s already gaining traction with larger enterprises because of its detailed business intelligence (BI) offerings. 

“Decision intelligence is a new category of software that facilitates the commercial application of artificial intelligence, providing predictive insight and recommended actions to users,” Gilroy said. “It is outcome focused, meaning a solution must deliver against a business need before it can be classed as DI.

“Recognized by Gartner and IDC, it has the potential to be the biggest software category in the world and is already being utilized by businesses across a variety of use cases, from personalizing shopper experiences to streamlining complex supply chains. Brands such as Nike, PepsiCo, and ASOS are known to be using DI already.”

Read next: Top Performing Artificial Intelligence Companies