AI and Machine Learning:
Artificial intelligence and machine learning have no doubts been a trending topic, with these technologies finding applications in everything from quantum computers and medical diagnostic systems to consumer electronics and smart personal assistants.
According to the IDC predictions, the global revenue from AI hardware, software and services will reach $156.5 billion this year, which is 12.3% more compared to 2019.
However, there is a risk of not seeing the whole picture when it comes to trends in the development and use of AI and machine learning. CRN / USA offers its insight into five key trends in this area - not only in terms of new applications but also approaches to its development and use.
Hyper-Automation: AI, ML & RPA
Hyper-Automation is the IT megatrend that Gartner is pointing out and it means that anything that can be automated will be done so. For example, let’s take legacy business processes. The pandemic has accelerated the implementation of this concept, also known as digital, or intelligent, process automation.
AI and machine learning (ML), along with robotic process automation (RPA), are the key components and major driving force of hyper-automation. To be successful, hyper-automation projects cannot rely on rigid software packages: automated business processes must be able to adapt to changing circumstances and respond to unforeseen situations.
This is where AI, ML models and “deep learning” come to the rescue, using so-called “training” algorithms, as well as data generated by an automated system, in order to automatically respond to changing business processes and requirements and do it more and more successfully. Deep learning is a subset of ML that uses neural network algorithms to train on large amounts of data.
AI Engineering as a Development Strategy
According to the Gartner research, only about 53% of all AI projects successfully progress from prototype to implementation. Many organizations are faced with serviceability, scalability, and manageability issues as they strive to implement newly developed AI systems and machine learning models, and such projects often fail to deliver the expected return.
In its list of top strategic IT trends of 2021, Gartner points out that organizations are now realizing they need a clear AI engineering strategy that will improve “performance, scalability, interpretability, and robustness of AI models” while delivering “full return on investment in AI”.
Gartner also notes that it is necessary to have a well-structured AI engineering process that includes the elements of DataOps, ModelOps and DevOps so that AI development can become a part of the entire DevOps process, rather than a collection of isolated specialized projects.
Growing Use of AI and ML in Cybersecurity
AI and machine learning technologies are increasingly used in cybersecurity systems, not only corporate but consumer. The developers of these systems are in a perpetual race, relentlessly updating and improving their systems to keep up with the ever-changing landscape of cyber threats: malware, ransomware, DDoS attacks, and more. AI and machine learning technologies can help identify already known threats and their new modifications.
AI cybersecurity tools can collect data from internal transaction processing systems, communication networks, digital activity and websites data, as well as external public sources, and use AI algorithms to recognize patterns and identify dangerous activity - detecting suspicious IP addresses and potential data theft.
According to IHS Markit, in the consumer segment, AI is mainly focused on home security systems that are integrated with CCTV cameras and voice-assisted burglar alarms. But IHS also states that the use of AI will expand with the prospect of creating a smart home, where the system learns the habits and preferences of the inhabitants, which will allow it to identify intruders more reliably.
When AI / ML Meets IoT
The Internet of Things has been a rapidly growing market in recent years, and, as Transforma Insights predicts, the global IoT market will grow to 24.1 billion devices by 2030, generating $1.5 trillion in revenue. The development of IoT is becoming more and more intertwined with the use of AI and machine learning.
These technologies, along with deep learning, are already helping IoT devices and services become smarter and more secure. But the advantages are mutual: the larger the amount of data AI and ML use, the more successfully they work - and that is what networks of IoT sensors and devices are ready to provide.
In an industrial setting, plant-wide IoT networks can collect process and operational data, which is then analyzed by AI systems to improve the productivity and efficiency of manufacturing equipment and predict the need for maintenance. This concept has already been coined as "artificial intelligence of things" (AIoT) and can radically change the approach to industrial automation.
AI & ML in Business: Ethical Aspect of Usage
The Washington Post writes that back in the middle of the year, as protests against racial injustice in the United States peaked, Microsoft, IBM and Amazon announced that they would restrict the use of their AI facial recognition technology by the police until federal laws governing the use of this technology were passed.
This has led to a growing concern about the ethical aspects of the increasing use of AI technologies. It will suffice to mention the use of artificial intelligence for cyberattacks and disinformation using deep fake technology. But there are also more ambiguous areas: for example, the use of AI by law enforcement agencies for video surveillance, as well as by commercial organizations for marketing and customer interaction. And this is not to mention the more profound aspects of the possible applications of AI in systems that can completely replace humans.
In one of their articles dated December 2019, Forbes states that the first step is to ask the necessary questions, and it is already being done. Certain areas of application may require federal regulation and legislation, for example, the use of AI technology by law enforcement agencies.
In the business arena, Gartner recommends creating independent AI ethics councils to avoid mistakes that could jeopardize a company's brand, lead to government regulation, or even "boycott and crash the business." Such councils, which include client representatives, can advise on the possible implications of AI development projects, increasing their transparency and accountability.