IBM’s announcements offer capabilities to promote more transparency across AI use cases to reduce any potential resistance to automating business processes.
IBM today showcased additional artificial intelligence (AI) capabilities that include an AI Factsheets capability that promises to make it simpler to understand and track the provenance of AI models. Those factsheets are intended for artificial intelligence models to be roughly the equivalent of nutrition labels for foods or information sheets for appliances.
There is also now a more efficient intent classification model for Watson Assistant that makes it simpler to train the artificial intelligence tool using less data. That capability is significant because many processes that IT organizations are trying to automate don’t have the amount of data that previously would have been required to train a virtual assistant.
IBM also previewed today a reading comprehension capability that enables the IBM Watson AI platform to attach confidence scores to answers to questions that will also now be more precise because of forthcoming advances made to the natural language processing (NLP) capabilities of IBM Watson Discovery that are currently in beta. Another NLP capability in beta showcased by IBM makes it easier to extract question and answer pairs from frequent accessed question (FAQ) documents.
Finally, IBM Watson Discovery has also been updated to add support for 10 additional languages, including Bosnian, Croatian, Danish, Finnish, Hebrew, Hindi, Norwegian (Bokmål), Norwegian (Nynorsk), Serbian, and Swedish.
Artificial Intelligence drives digital transformation
As part of an overall effort to accelerate digital business transformation initiatives in the wake of the economic downturn brought on by the COVID-19 pandemic, AI models are starting to be employed more widely. For example, EY, a global leader in consulting services, assurance, tax, strategy, and transactions has leveraged IBM AI technologies to launch Diligence Edge, which includes a custom NLP model trained on EY proprietary merger and acquisition language. Other organizations that are making use of IBM AI technologies include Regions Bank, Japan Airlines, and NatWest.
In general, IBM is trying to promote more transparency across various use cases involving AI in part to reduce any potential resistance to automating additional business processes. The AI Factsheets capability will also make it easier for governments to create and enforce governance requirements for AI applications, says Aleksandra Mojsilovic, head of AI Foundations at IBM Research.
“We are sharing with organizations a toolkit for meet those requirements,” says Mojsilovic.
In general, organizations are moving ahead with investments in AI but there are concerns they might be building applications that run afoul of regulations that are still being crafted by most countries. The depth of understanding of what is and what is not possible to achieve with AI among lawmakers around the world is limited.
In the meantime, AI models coupled with other leading-edge technologies such as robotic process automation have the potential to modernize a wide range of antiquated business processes that are adversely impacting customer satisfaction in what is rapidly becoming a new digital age. Organizations that don’t make those investments will soon find themselves at a significant competitive disadvantage.
The challenge, of course, is determining where to apply AI technologies today in anticipation of what additional advances will soon be made available tomorrow.