Introducing AI’s lost twin, engineered intelligence

We are on the brink of a fourth AI winter as faith that AI will create tangible value that is justified by its costs begins to waver.

As articles from Goldman Sachs and other research institutions fall like leaves, there is still time to thwart the next AI winter, and the answers have been right in front of us for years.

Something is missing

In most scientific fields, breakthroughs occur in the laboratory and are then passed on to engineers for real-world applications.

When a team of chemical researchers discovers a new way to form adhesive bonds, the discovery is passed on to chemical engineers to design a product or solution.

The breakthrough inventions by mechanical physicists are then handed over to mechanical engineers to design a solution.

However, when breakthroughs are made in AI, and there is no clear field for applied artificial intelligence, organizations invest in hiring PhD data scientists with the ambition to make scientific advances in the field of AI, and instead try to design real-world solutions.

The result? 87% of AI projects fail.

The emergence of artificial intelligence

“Engineering Intelligence” (present participle: “intelligence engineering”) is an emerging field focused on real-world applications of AI research rooted in engineering – leveraging scientific breakthroughs and raw materials to design and build safe, practical value. It empowers domain experts, scientists, and engineers to create intelligent solutions without having to be data scientists.

Leading industrial organizations are beginning to restructure their research-to-engineering pipelines, forming new partnerships with academia and technology vendors, and creating the ecosystem conditions in which AI research can be handed off to intelligence engineers in the same way that chemical research is shared with chemical engineers.

result?

Breakthrough applications that create value, are actionable, and are concrete use cases that data scientists and technology vendors would never have discovered based on the data alone.

Five steps to adopting intelligence engineering in your organization

Expertise is at the heart of intelligence engineering and is expressed as skills. A skill is a unit of specialized knowledge acquired through practice. Theory and training can accelerate the acquisition of skills, but skills (and therefore expertise) cannot be acquired without practical experience. Assuming your organization already has experts, here are five practical steps you can take to adopt the discipline of intelligence engineering and where you deviate from traditional approaches to leverage AI:

Traditional approaches to adopting AI (which has an 87% failure rate) are:

1. Make a list of your issues.

or

1. Examine the data.
2. Select a set of potential use cases.
3. Analyze use cases for return on investment (ROI), feasibility, cost, and timeline.
4. Choose a subset of use cases and invest in executing them.

The intelligence engineering approach to deploying engineered intelligence is as follows:

1. Create a heatmap of expertise across existing processes.
2. Evaluate which expertise is most valuable to your organization and assess how abundant or scarce that expertise is.
3. Select the top five most valuable and scarce areas of expertise within your organization.
4. Analyze ROI, feasibility, costs and timelines to design intelligent solutions.
5. Select a subset of worthwhile cases and invest in their implementation.

Creating a new wave of value with AI

As intelligence engineering is introduced into organizations and intuitive applications are developed and operationalized, this new capability can be leveraged to go beyond existing expertise to create new opportunities to engineer secure, actionable value across the organization and ecosystem.

As organizations, industries, and educational institutions build out their artificial intelligence programs, organizations, individuals, and society will benefit from AI’s previously unrealized economic and social potential, creating new kinds of jobs and ushering in a new wave of value creation.

Brian Evergreen is author of Autonomous Transformation: Creating a More Human Future in the Age of Artificial Intelligence.

Kent Anderson is the author of Designing Autonomous AI.