MarketsandMarkets has done a recent research on AI Market. TimesTech spoke with Sushmit Chakraborty, Senior analyst – ICT, MarketsandMarkets to know more about the research and its outcome.
Read the details here:
TimesTech: How AI as a technology has developed and matured?
Sushmit: AI as a technology in the past decade (2010-2020), has made significant strides in research and development, with the evolution of deep learning models ushering in unprecedented growth across real-time applications in the field of cognitive learning to intelligent analytics.
AI models are typically categorized into two classes: weak AI and strong AI. Weak AI comprises capabilities that mimic human cognitive functions by automating time-consuming tasks such as deciphering voice or text or recognizing objects in a photo or video stream. Strong AI combines weak AI capabilities into a construct replicating human intelligence in terms of cognition, reasoning, and consciousness. While researchers across academia and technology development continue to pursue strong AI, weak AI capabilities are widely available today that can be incorporated within applications and processes to enable organizations in achieving digital transformation across business functions.
TimesTech: You recently researched the AI market. Tell us more about the research.
Sushmit: The report provides a holistic view of the artificial intelligence market and covers the following research objectives:
- Evaluate current and future opportunities in the AI market by offering (hardware, software and services), technology (machine learning, natural language processing, context awareness, computer vision), deployment mode (cloud, on-premises), organization size (small and medium enterprises, large enterprises), business function (finance, security, human resources, law, marketing & sales, other business functions), vertical (banking, financial services and insurance, retail and ecommerce, automotive, transportation and logistics, government and defense, healthcare and lifesciences, IT/ITes, energy and utilities, telecommunication, manufacturing and other verticals) and region (North America, Europe, APAC, Latin America and the Middle East and Africa)
- Provide a piece of detailed information related to qualitative factors (drivers, restraints, trends and opportunities) influencing the market growth
- Assess the competitive landscape for stakeholders and market players, profile key players and analyze the market share/rankings and core competencies
- Analyze competitive developments in the vendor landscape such as partnerships, collaborations, new product launches, and mergers and acquisitions in the market
TimesTech: What are the challenges and opportunities in tapping the AI market?
Sushmit: Challenges:
- Limited number of AI experts and shortage of skill-set in terms of workforce developing, managing and implementing AI systems
- Data security and privacy associated with managing a substantial volume of data from diverse sources. Most AI models, in order to increase their efficacy, need to be trained on real-time datasets which may result in breach of sensitive and confidential information
Opportunities:
- Growth in ethical AI governance framework and guidelines to reduce discrimination biases around ML algorithms will enhance trust among AI users and boost the adoption of AI-based technology solution development
- Government-focused initiatives to facilitate the adoption and investment in AI technology will lead to the creation of new revenue-generation opportunities for market stakeholders
TimesTech: How AI is making automotive more capable than ever, and what are some new techs that we can see in the future?
Sushmit: Integration of AI in automotive has come a long way, specifically in ensuring the safety of drivers and passengers. Automotive OEMs are increasingly investing in ADAS that can detect and respond to potential hazards on the road. The data collected by the cameras, sensors, LiDARS, and Radars are processed in real-time through AI algorithms to make informed decisions. Below are some of the key use cases for AI in automotive:
- AI technology has helped to reduce the number of accidents on the road by providing drivers with real-time information about potential hazards or automatically sending alerts to emergency response services in the event of an accident.
- Advanced AI-power navigation systems make it easier for drivers to make better navigation decisions as these systems can suggest better routes. This is done by an AI-based system that gathers and analyses millions of data points on information about nearby road closures, traffic jams, accidents, construction work, and road conditions.
- AI technology helps predict a potential component failure in the automobile by monitoring and analyzing hundreds of sensor data points.
- AI technology is also used in automobiles to analyze driver behavior that can cause traffic hazards.
- AI is also crucial in the development of electric vehicles, specifically in the battery development process. Engineers are using AI-based algorithms to predict battery performance under various conditions, such as road and weather conditions.
A lot of developments are happening in the Automotive sector, and we could witness the introduction of innovative technologies/solutions supplementing AI adoption.
- Biometric seats: Biometric Seat are integrated with highly reactive sensors which can monitor the driver’s breathing rate, heart rate, and body temperature. AI technology could help analyze sensor data and send warnings if the driver is too stressed or tense.
- Automotive liftgates: Most automotive OEMs have started using this feature to enhance customer experience. It allows the user to automatically open the rear liftgate by waving a hand.
- AR-based windshield displays: This technology is being used in vehicles to display vehicle information such as current weather, gas mileage, speed, blind spots, etc., enabling driver to follow directions without the need to take their eyes off the road.
- Firmware Over the Air (FOTA) Update: This technology enables firmware downloads and updates for specific electronic control units (ECUs) inside a car remotely.
TimesTech: What are the challenges in making vehicles fully autonomous using AI?
Sushmit: A lot of investment in autonomous vehicles is being made to make it a real possibility in some parts of the industry, such as agriculture, transportation, military and defense, etc. Several automotive OEMs are developing comprehensive ADAS platform that combines features and capabilities from multiple platforms, such as cameras, RADARs, sensors, and LiDAR. To realize Level 5 vehicle autonomy, engineers need to develop machine learning-based technology for integrating AI into the model. However, there are several challenges in making fully autonomous vehicles using AI.
There are no widely accepted Machine learning algorithms to train AI models that could assure a more precise machine learning training process, which would be able to detect and classify objects or things more efficiently than a human driver.
- AI technology is not as adaptable as the human brain and fails to subtle changes. For instance, pedestrians are relatively unpredictable. Engineers are facing challenges in making reliable algorithms that can manage and deploy a system that can support not only 100% of what is predictable but also the small fraction that is unpredictable.
- Implementation of AI in autonomous vehicles leads to cybersecurity risks. AI systems can be easily misguided by simply pasting a sticker on the road sign to prevent its recognition. Therefore, it is quite challenging to develop AI security policies for the development and deployment of AI functionalities, which is mutually agreed upon and accepted by the supply chain members in the automotive industry, including Automotive OEMs, Government organizations, software developers, AI system developers.
- Localization is also one of the major challenges for AI system developers in the automotive industry as they need to train the AI models keeping in mind several factors such as different languages, cultures, and demographics.
