AI is being democratized in all areas of business with the challenge of ensuring that it is "reliable", according to an IBM expert

VALENCIA, 19 Oct. (EUROPA PRESS) -

The Data

This is what he indicated during the lecture 'AI as a tool of transformation', which he gave in Valencia as part of a day organized by Seidor and the Business and Humanism Chair of the University of Valencia (UV) on the occasion of the presentation of the II Barometer of the Digital Transformation in the Companies of the Valencian Community

Blas Canet has explained that AI can improve the decision-making process in real time and "unload employees from unproductive and low-value work", to the point where "solutions can be applied in any area of ​​the company." "artificial intelligence. In that sense, he explained that in 2017, 20% of companies used artificial intelligence and that in 2022 it would be 50%, more than double.

What has changed, according to Canet, is that before AI models such as 'machine learning', 'deep learning' or generative were prioritized "which were specific for a purpose and only served for that", and also required "very expert people and it was difficult to find talent in the market at a reasonable price.

Now, these models "cohabit" with the so-called foundational models, which are "general purpose" and can be modified for different uses. "This is an accelerator, it allows this type of model to be implemented much more quickly and economically, without having to recruit talents that I may not be able to find within a time frame or at a cost."

Thus, while "recently companies have applied AI to specific points and very specific processes in their business areas, the trend is for AI to be infused into all business processes and to be a key part and addition to their work philosophy." ". We have gone "from working on very specific cases, for example customer classification or predictions to adapt supply to demand, to new, much more diverse use cases."

Canet has assessed that now "each artificial intelligence journey is different" and "it is a technology that is being applied in all business areas." For example, she has noted the “incredible opportunities around enterprise data.” She has also specified that in human resources resumes can be analyzed, abandonments can be detected, meetings and work can be transcribed or used in marketing for segmentation and prediction of demand.

Throughout this framework, Canet has stressed the importance of "making AI trustworthy" because its use "entails great benefits and can entail great risks." As an example, he has pointed out that in the financial sector there could be biases based on gender, social origin or discrimination based on place of residence when managing credits. "Bias must be managed because there can be very unfair models," she warned.

Another issue raised is the ownership of the data that is managed or protection so that it cannot be used by third parties.

In this sense, he has defended that AI management must "ensure responsible AI" in its use, framed in regulatory and transparency policies and with "explainability", in reference to the ability for users to understand its operation and how you achieve your results.