With all the hype surrounding Generation AI, it’s no surprise that it’s the dominant enterprise AI solution, according to a Gartner survey published in May. Twenty-nine percent of 644 executives at companies in the U.S., Germany, and the U.K. said they already use Generation AI, and it’s more widely used than other AI-related technologies such as optimization algorithms, rule-based systems, natural language processing, and other types of ML.

The real challenge, however, is to “prove and value” the value of projects, not only in terms of total cost of ownership and the broad benefits they will deliver, but also in terms of obstacles such as a lack of confidence in the technical aspects of AI and difficulties in providing sufficient amounts of data. However, these are not insurmountable challenges.

Data protection

The first step in AI and Gen AI projects is always to get the right data. “In cases where privacy is important, we try to anonymize as much as possible and then proceed with training the model,” says Vincenzo Laveglia, a technologist at the University of Florence. “A balance between privacy and utility is needed. If the information content of the data remains the same after anonymization, it is still useful. But as soon as personal or sensitive references are removed and the data is no longer useful, a problem arises. Synthetic data avoids these difficulties but is not exempt from the need for a compromise. We need to make sure there is a balance between different classes of information, otherwise the model becomes an expert on one subject and a very uncertain one on others.”