OpenAI has announced initial support for plugins in ChatGPT, its widely known artificial intelligence model that produces human-like text from a basis of deep learning. The plugins will help ChatGPT access a range of applications and perform various actions for end users of applications.
OpenAI’s documentation page highlights various use cases:
- Retrieving real-time, as-it-happens information, such as scores and stock prices.
- Accessing standing information, such as company documents.
- Performing services for the user, such as ordering food or making reservations.
Building Use Cases
Plugin developers, OpenAI said, will “expose one or more API endpoints, accompanied by a standardized manifest file and an OpenAPI specification. These define the plugin’s functionality, allowing ChatGPT to consume the files and make calls to the developer-defined APIs.”
The use cases OpenAI mentions, such as pulling sports scores and flight reservations, are some of the easier ones to imagine. The documentation page goes on to note that “over time, we anticipate the system will evolve to accommodate more advanced use cases.”
That’s where things will get interesting.
The Coming Transformation
Although the bulk of the early attention on ChatGPT has centered on the question of who’s going to be writing what (and how will we know?), much of its potential for disruption and transformation has resided elsewhere.
Search functions are likely to change radically. Curation will, too. How people make their choices across a range of consumer options will be profoundly affected, as will choices about what payment vehicle to use. Those methods themselves are in for changes, too. AI will eventually make the payment choice for consumers, in the background, based on what’s most advantageous for the buyers.
Those things, cautioned Marco Salazar, Javelin Strategy & Research Director of Tech and Infrastructure, “are still a few steps away.” But they’re coming, he said.
“ChatGPT is becoming a really big testing ground,” Salazar said. “It’s at scale and at speed.”
Salazar’s recent report Open Banking Pushes Interoperability to the Payments Forefront highlighted what’s in store for AI tools as they absorb data to train the models. OpenAI’s call for plugin development opens the floodgates, he said.
In banking, the coming changes augur in favor of customers. In the Javelin report 2023 Digital Banking Trends & Predictions, authors Mark Schwanhausser and Emmett Higdon noted that automation and artificial intelligence would push financial institutions further along the path of reduced friction and heightened engagement through enhanced transactional, educational, and analytical customer experiences.
Effective Anti-Fraud Tools
FIs are already leveraging artificial intelligence and machine learning to know who’s on the other end of an interaction or transaction with greater accuracy. In an October 2022 Javelin report, Data Detective: Advancements in Contextual Intelligence, Senior Analyst Suzanne Sando detailed the challenge, writing:
“Because it’s nearly impossible to find a consumer without some semblance of a digital footprint, cybercriminals have mountains of data available to them for exploitation” in gaining unauthorized access to accounts.
AI and machine learning help blunt those attempts by zeroing in on characteristics—keystroke cadence, mouse control, browser choice, etc.—that help differentiate legitimate users from bogus ones. Imagine the possibilities for when AI becomes more robust and more embedded into everyday interactions.