Tech has no shortage of buzzy new technologies – and cutting through the hype to see what will actually impact the enterprise can be challenging. We're here to help. Starting in 2021, our contributors will give a rundown on an emerging tech and whether or not it'll pay off to pay attention to it. For artificial intelligence in 2021, here’s our look at formative AI.
What is Formative AI?
Formative AI is an umbrella term that refers to artificial intelligence that changes dynamically in response to a situational variance — as the NYC Media Lab puts it, it’s “AI that ‘creates’ things.” There are multiple types of formative AI. Generative AI is a type of formative AI that creates new content or alters existing content — for example, in deep fake images and videos. And formative AI models can be designed to evolve and adapt over time. The applications of formative AI are as varied as synthetic data generation to train machine learning models and AI artwork or deep fakes made with nefarious intent.
How Long Has It Been Around?
It’s hard to pin a date of birth on formative AI because it’s an umbrella term for a variety of technologies, many of which are emerging and not yet fully established in the enterprise.
Gartner identified several technologies under the formative AI umbrella as set to become productive over the next two to ten years. Some of those technologies, like augmented AI design, are just emerging but are already in use in some cases and predicted to be an emerging category to watch.
Why Are People Paying Attention to It Now?
Gartner placed formative AI on its list of must-watch technologies for its 2020 hype cycle report, citing its potential for disruption and widespread possible enterprise applications. Formative AI is a means by which the reams of data that organizations are increasingly gathering, generating and storing becomes more useful.
Gartner identified a few formative AI technologies that enterprises should pay attention to, including AI-augmented design and development, adaptive machine learning and generative AI. These tools have potential for efficiency, productivity, creativity, prediction and cost-reduction for businesses. Not every technology is a fit for every business, however — it’s a mistake to fall prey to the buzz and jump on a tech without understanding how it makes sense for operations.
Formative AI also has the potential to reduce the costs of machine learning for enterprises, making it extremely attractive to cost-conscious IT pros.
Who Benefits From It?
The big winners could be budget-conscious, data-processing businesses. Expect to see it roll out across medical operations, predictive professions and the enterprise departments tasked with maintaining or improving internal operations.
Adaptive machine learning could result in better outcomes with fewer time and resources — for example, diagnostics that combine machine learning with expert knowledge and are more accessible than ever before as powerful computers become less expensive.
Or generative AI can play a role when the needed data for good predictions or analysis isn’t available by generating data to meet particular conditions. This generated data can be used to train learning models, for example, or to avoid issues of privacy (as with medical or financial data). It can also be used to create data that is more accurate or more free of bias, improving learning models and algorithms.
AI-augmented development also has the potential to help organizations overcome staffing and training concerns that stymie AI adoption or scaling. With artificial intelligence available to assist designers and developers, errors can be spotted more effectively and simple programming can be automated.
Where Can You Get It?
There are applications on the market now that fit under the formative AI umbrella — for example, Kite’s AI code completion software or Atomwise’s use of AI in drug discovery. Generative AI in the form of deep fakes has already been put to use in the entertainment industry and as art or educational projects. Augmented AI design is already at play in CX products but is a category with wide-ranging potential.
Other tech in the category remains in the R&D or early development stages, like self-supervised AI.