Evolution from Narrow AI to General Purpose AI
When most people think of Artificial Intelligence (AI) today, they think of what is referred to as “Weak AI”, which is limited in terms of the type and variety of problems that may be solved. In other words, today’s AI is largely purpose-built for specific use cases, industries, and scenario-specific problem solving. In contrast, “Strong AI” provides the ability to solve a broad based of problems that are cross-domain, spanning many industries, companies, and solutions.
Strong AI represents and evolution of today’s silo-oriented AI solutions in which intelligence becomes more general purpose. The terms of often used to describe this next generation AI are Artificial General Intelligence (AGI) or simply general-purpose AI. The term “general” is used to describe this evolved version of AI as AGI is anticipated to be a more generally available and useful function across a variety of applications, products, and services.
An example of weak AI is Amazon Alexa, which is typically used for single-purpose functions such as playing music, answering questions, providing a timer, ordering from Amazon, etc. These functions do not necessarily learn from one another (yet), and furthermore, the user interface (voice) is the only way (today) that Alexa can interface with the user. Alexa does not listen into conversations to learn about habits and apply towards end-user wants and needs, which would be more like what strong AI would do.
Furthermore, the software for Alexa is written to support Amazon’s business interests. It is not (yet) written to work on a cross-industry, inter-company basis. However, Alexa is finding its way into more end-user interfaces such automotive vehicles, which could be a more natural way of leveraging AI to learn from user habits such as telling Alexa to take notes while one is driving. Perhaps a better example of AI that is seamless is simply Google’s Gmail, which represents some steps towards stronger AI as it recognizes one’s interests and presents information accordingly.
General Purpose AI Converges with Data Analytics and the Internet of Things
The Internet of Things (IoT) is rapidly expanding, and much of the data generated is machine-generated and of the unstructured variety, requiring big data analytics tools. As useful as big data tools are for managing data, they benefit significantly from the introduction of AI algorithms to support analytics solutions. AI will provide improved decision-making capabilities for IoT as well as scalability as the mountains of data continue to accumulate from IoT networks and systems.
AI coupled with advanced big data analytics provides the ability to make raw data meaningful and useful as information for decision-making purposes. The use of AI for decision making in IoT and data analytics will be crucial for efficient and effective decision making, especially in the area of streaming data and real-time analytics associated with edge computing networks.
Whereas big data analytics facilitates the ability to transform otherwise meaningless data into information, AI adds the decision layer, enabling information to become actionable insights. In fact, the use of AI for decision-making will become the norm for IoT-related data analytics, especially as data processing becomes more rapid and autonomous as computing is pushed to the edge of networks, closer to the point of application and service engagement.
The role and importance of AGI in IoT data analytics cannot be overstated as general-purpose AI provides the ability to analyze information on a cross-platform, industry, and company basis. This is anticipated to start in federated manner, but ultimately to evolve to a flatter network structure as AI peering (e.g. AI algorithms from different sources communicating) becomes the predominant means of applying AGI to problem solving.
General Purpose AI will provide a Utility Function, acting as a Resource to virtually any other Device or Network Element such IoT Platforms and Software Programs via APIs
The combination of AI and IoT represents an AIoT market opportunity that will transform more than just IoT data processing and decision making. AIoT will transform entire industries. The reality today is that AIoT is largely a silo capability (e.g. weak AI), largely relegated to consumer-oriented products such as appliances and electronics.
This is anticipated to rapidly change, however, as leading companies such as SAS are working on solutions that provide more generalized value from a broad array of data, regardless of the application, product/service provider, or specific product or service.
Taken together, AIoT with advanced big data analytics, will enable decisions as a service, which is a natural evolution of the data as a service market. In other words, the “as a service” model provided by leading companies such as Amazon, Google, IBM, and Microsoft will provide more than just data – these services will provide actionable information, and in some cases, autonomous decisions.
Learn more about AIoT at the AI Time Journal AIoT Trend Talk
The volume human data involved in IoT is minuscule compared to the amount of machine-generated data, which is also fast moving and typically streaming in nature. This amplifies the need for AGI solutions that can work with big data analytics to utilize vast amount of data and make decisions quickly.
As part of its AI research and strategic advisory practice, Mind Commerce evaluates the general purpose AI market including leading companies, services potential, technology integration, and application ecosystem. Our research analyzes the AI agent market and the relationship between general purpose AI with other technologies including edge computing, 5G networks, and blockchain.
General Purpose AI in support of Medical Devices alone represents a $1.4B Market Globally by 2023
The aggregate AGI market is anticipated to exceed $50 billion globally by 2023. While hardware remains the largest market component during this period, AGI services and software are the fastest growing elements at 56.5% CAGR and 51.4% CAGR respectively through 2023. It is anticipated that edge computing, 5G, and future 6G technology market infrastructure and services will be the biggest telecommunication related beneficiaries for AGI.
AGI capabilities in support of the AI chipset market represents a $542 million global opportunity by 2023. Key AI capability areas supported include deep learning, supervised, unsupervised, and reinforcement learning. It is important to note, however, that AI chipsets largely represent weak AI today, but in the future will become fully integrated with robust AGI platforms and systems in the future.
The healthcare sector will be a leading industry vertical exceeding $8B billion by 2023 with AGI supported medical devices alone exceeding $1B globally. North America is anticipated to be the largest regional market for AGI through 2023, exceeding $17 billion. AGI will be particularly useful for the market for quantified-self in healthcare, as general purpose AI will assist end-users in diagnosis and overall self-help wellness.
AGI will add significant value to many other industries sectors such as agriculture in which it will support precision farming, livestock monitoring, drone analytics, agriculture robots, irrigation, insect management, and many other important agriculture technology functions. Mind Commerce is open to dialog about AI evolution towards general purpose functionality. Contact us to schedule a briefing.
Technology, Media, and Telecom (TMT) sector professional with 30+ years developing, managing, and advising about networks, devices, apps and services. This includes planning, engineering, product management, business and corporate development.
Skilled in system design, network architecture, and application development. Experience areas include digital identity, presence and location, SS7 and SIP signaling, network, cloud, and end-point intelligence. Range of experience spans fixed and wireless network operators, service bureau, and application provider companies. Solutions designed and implemented include location services, wireless data infrastructure, mobile messaging, content management, prepaid and stored value applications, and mobile marketing.
Author of many technical papers about various telecommunications subjects including the published reports “Yes 2 Prepay” and “Data on SS7” as well as co-author of the books “Wireless Intelligent Networking” and “Mobile Positioning and Location Management”.
Thought leader in emerging and disintermediating technologies including 5G, 6G, and Artificial Intelligence of Things (AIoT). Additional areas include Broadband, Cloud Computing, Data Management Analytics, Edge Computing, Immersive Technologies (Augmented and Virtual Reality), Industrial Automation, Internet of Things, and Robotics.