What is augmented intelligence?

How does augmented intelligence work?

Augmented intelligence uses a mix of different technologies to help humans in their tasks. For example, tools that aid humans in a task but don't involve AI, like data management software, are still considered to be augmented intelligence.

AI-enabled technologies commonly used in augmented intelligence, however, include machine learning, deep learning, machine vision and natural language processing (NLP).

Machine learning and AI systems are typically trained using a host of different learning methods -- like supervised, unsupervised, semisupervised and reinforcement learning -- that all require large amounts of data to train. The same idea applies to training a machine vision system to recognize images, or training an NLP system to understand human language and text. These AI-enabled systems are trained to complete or automate a specific task and, in some cases, take the place of a human role. As an example, an AI chatbot can be designed to handle customer interactions, taking the place of a human agent.

Unlike traditional AI systems, instead of replacing or completely automating a task, augmented intelligence systems are instead used to aid a human in a task. For example, the earlier mentioned AI chatbot could be used to initially answer a customer service request, ask the customer a series of questions related to their query, and then direct the customer to the appropriate human agent. Augmented intelligence systems can also do the following:

• Gathering structured and unstructured data.
• Analyzing and interpreting data.
• Providing actionable data.
• Streamlining workflows by automating smaller, repetitive tasks.

How is augmented intelligence different from AI?

Augmented intelligence involves the use of any technology to enhance human intelligence, while AI is defined as the simulation of human intelligence by machines. On the surface, the terms seem different, but in practice, they are highly complementary since, as noted, many AI applications are used to augment human intelligence rather than replace it. Additionally, augmented intelligence includes tools that improve our ability to think, but do not necessarily involve AI, such as software that links to content, manages data or provides a new user interface.

At the same time, there are AI systems built to replace tasks traditionally done by humans. But for the foreseeable future, these systems will continue to augment other human roles that involve things AI can't do, such as provide oversight, add context or manage risks. An example of this is the effort to build fully self-driving cars. While today's autonomous vehicles have made substantial progress in operating autonomously, even the most sophisticated self-driving taxis must be overseen by humans in special control rooms to respond to unforeseen issues.

Augmented intelligence vs. artificial intelligence in the enterprise

In business discussions, it can be useful to characterize augmented intelligence as a spectrum. At one end, there is no augmentation, such as a human writing an article by hand. Further along the continuum, a word processor might facilitate easier writing and editing. In the middle, humans might work with generative AI (GenAI) prompts to automate much of the process and then edit the results. At the end of the spectrum, fully autonomous AI systems might identify content needs, craft the content and automatically publish the results.

Few people are advocating for full AI autonomy today -- whether in researching, writing, driving cars, diagnosing and caring for patients, manufacturing, or running a business -- due to various safety, governance, risk management and reputational implications. However, these more fully autonomous AI systems are one of the goals of artificial general intelligence, which aims to simulate how humans learn and think.

Frank discussions about augmented intelligence sharpen business strategy

Because augmented intelligence is a term broadly applied to any technology that assists a human's ability to reason, make decisions and perform tasks, it is not particularly useful to think about on its own. However, when the use of augmented intelligence is discussed frankly in the context of specific business initiatives and disciplines, it can sharpen AI strategies and calm employee anxiety about AI replacing jobs.

• In governance, risk and compliance management, for example, discussions about the use of AI augmentation can help frame best practices for managing AI risks.
• In digital transformation, discussions about the use of augmented intelligence can ease the fear of layoffs and support efforts to make jobs more meaningful and productive.
• In user experience and human-computer interaction, discussions about augmented intelligence can help identify opportunities to improve the coordination among humans, AI and other systems that enhance human experience and safety while driving efficiency.

History of augmented intelligence

Augmented intelligence and AI have a rich history in research and business. The following timeline reflects the evolution of and tension between the two approaches, and focuses on how augmented intelligence has figured into business, policy and technology discussions.

1955. John McCarthy and other early AI luminaries coin the term artificial intelligence to promote a 1956 summer workshop at Dartmouth College convened to explore mechanisms for enhancing and rivaling human intelligence. The tension between the two competing approaches around rivaling human intelligence versus augmenting it has continued to this day.

1956. William Ross Ashby discusses the term amplifying intelligence in the book Introduction to Cybernetics, which explored innovations in information theory, operations research and human-computer interaction.

1960. J.C.R. Licklider explores human-computer interaction challenges and opportunities in a paper titled "Man-Computer Symbiosis."

1962. Douglas Engelbart discusses a framework for augmenting human intelligence. He described augmentation as "increasing the capability of a man to approach a complex problem situation, to gain comprehension to suit his particular needs, and to derive solutions to problems."

1985. Howard Rheingold introduces "mind amplifier" in Tools for Thought, a popular take on technologies that enhance human cognitive abilities.

1987-1994. Overinvestment and failure to deliver leads to a second AI winter and a decline in the use of the terms AI and augmented intelligence.

1990. Steve Jobs frames a computer as "like a bicycle for our minds" to describe the value of cheap computers in augmenting human intelligence.

2015. New AI algorithms and more scalable development techniques reignite enthusiasm and investment in AI.

2017. Gartner analysts Rita Sallam, Cindi Howson and Carlie Idoine introduce the term augmented analytics to describe AI's ability to democratize complex data analysis that previously required experts.

2019. Gartner predicts augmented intelligence will create $2.9 trillion of business value and 6.2 billion hours of worker productivity globally by 2021. The research firm defines augmented intelligence as a "human-centered partnership model of people and AI working together to enhance cognitive performance."

2019. Stanford University launches the Institute for Human-Centered Artificial Intelligence, which includes professors from engineering, medicine, social sciences and other disciplines. Its aim is to explore "advancing AI research, education, and policy to improve the human condition."

2020. AWS announces Amazon Augmented Artificial Intelligence to build and manage workflows that combine AI with human reviewers across its Mechanical Turk service, vendors and employees.

2022. American Medical Association President Jesse Ehrenfeld argues that the term augmented intelligence is a more fitting description of what machine intelligence provides in healthcare, emphasizing that the integration of human intelligence and "machine-derived outputs" aims to improve human health, not simply produce an output.

2023. Gartner expects AI-augmented development and the "augmented connected workforce" will be top trends in 2024, along with developments that will disrupt human jobs, such as AI-driven apps and democratized generative AI.

Present. AI continues to capture a much larger mind share thanks to the popularity of new GenAI services like ChatGPT, Google Gemini and Microsoft Copilot, which are raising new concerns about AI risks and job loss.

Benefits of augmented intelligence

Some of the benefits of augmented intelligence include the following:

• Enhances task completion time.
• Makes complex analysis and decision-making available to a larger audience.
• Can speed up application development processes.
• Can suggest next steps or input in a complex process.
• Can aid in predicting patterns.
• Can provide better feedback about issues.
• Can summarize complex information for easier understanding.
• Can sort through large volumes of data.

Challenges of augmented intelligence

Some of the challenges of augmented intelligence include the following:

• Reduces incentives for humans to learn fundamental skills.
• Shrinks requirement for entry-level roles that could later replace existing experts at retirement.
• Introduces new user experience design challenges that impede workflows.
• Accentuates complexities of processes that combine humans and AI.
• Heightens safety and risk management concerns without proper oversight.
• Collecting the required data to train a machine learning or AI system might involve ethical issues or become too costly.
• If the machine learning systems are trained on too little or not diverse enough data, then AI bias could be introduced into the system.
• The use of augmented intelligence might require personal and sensitive data -- such as in medical cases -- leading to concerns about how this data is stored and shared.

Applications of augmented intelligence

Augmented AI can be applied in various ways by organizations to improve efficiency, decision-making and customer experience. The following are some examples of augmented intelligence applications at work and in the world:

• Augmented analytics can democratize analytics and machine learning.
• Driver assistance systems enhance automotive and equipment safety.
• Augmented intelligence can identify bottlenecks and prioritize opportunities for improvement.
• Human resources departments can use augmented intelligence tools to tailor training programs to an employee's learning style and level of expertise.
• In healthcare, AI algorithms can help medical professionals diagnose illnesses and pinpoint treatments by analyzing large quantities of data.
• Sales and marketing can use AI to better understand customers, improve sales forecasting and customize marketing programs.
• Manufacturing plants can use augmented intelligence monitoring tools to facilitate repairs and maintenance of equipment.
• Voice assistants like Apple's Siri or Amazon's Alexa provide information to the user when asked and can also aid the user in accessibility-related tasks.
• In retail, augmented intelligence can help personalize customer experiences, and aid in optimizing workflows and identifying store bottlenecks.
• Financial businesses can use augmented intelligence to aid in processes that detect fraud and to help improve risk management.

Future of augmented intelligence

The terms augmented intelligence and AI will continue to evolve. Increasingly, enterprises will also use the term augmented AI to describe an emerging class of tools such as programming copilots, enterprise software assistants and new chat experiences.

Autonomous AI improvements will also shift the boundaries between augmented intelligence and autonomous AI. More autonomous apps might competently take on more tasks, but introduce new risks related to AI bias, hallucinations or data leakage that humans might have previously managed. An AI risk management strategy will be required to identify and develop processes to mitigate these risks.

Digital transformation efforts will increasingly need to strike the right balance between more autonomous AI and how this might reduce head count or make jobs less rewarding. Discussions about augmented intelligence, as noted, could help build trust across the organization and guide discussions that build buy-in to improve results for both the business and employees.

User experience design will continue to be an area of research as researchers, vendors, enterprises and employees innovate best practices for augmented intelligence experience design. This will include making user interfaces more interactive, reducing the number of unwanted prompts and integrating assistance more directly into the flow of work.

AI is a continually changing landscape. Learn more about how AI is implemented in enterprise settings.