The second day of the EMEA NewLaw Summit focused on tax. This is the other leg of PwC's division, Tax & Legal, which organizes the event. A continuation of the first day, which delved into the digital transformation experienced by the legal sector, with AI as a prominent factor. Artificial intelligence has not had such an accentuated protagonism in this case. But it has always hovered over the round tables and hallway conversations.
What's more, the conference that opened the day was expressly about generative AI. Alex Rayón, university professor and co-founder of the startup brainandcode.tech, was the one in charge of cheering up the audience in the first hour with an overwhelming speech. A review of the current situation of artificial intelligence that began by comparing it with biological intelligence. The first thing: a change in the concept. The speaker states that he prefers the definition of "digital intelligence" to "artificial". But that is a lost war as of today. So let's continue.
"In isolation we have been emulating human capabilities," Rayón notes, talking about the way artificial intelligence has traditionally been developed. A task was needed, such as making mathematical calculations or analyzing patterns in a table, and a specific algorithm was created. The speaker emphasizes the ability to discriminate data, which he considers to be the factor that has driven the improvement in human capabilities.
But it is also a factor that has promoted the best of AI capabilities. "People are capable of three forms of logical thinking: deductive, inductive and abductive," the university professor stresses. "Discriminative artificial intelligence, which is the historical one, was only deductive. It was able to take a lot of data and come to a specific conclusion. We would take an Excel and we would know which product was more easily purchased by a certain customer or we would take a large database of products and we could know which one was more likely to gain a good market share. In short, we would take a lot of data and come up with a particular piece of information."
However, most of the work of tax professionals is inductive and abductive, according to the speaker. "It's the ability to emerge an idea, singular and new. You get paid for it." That's why the new wave of AI brings changes for this field. What has the new generative era brought? Rayon answers: "The ability to submit in the inductive and abductive world. I can take the balance sheet of a profit and loss statement and I'm going to try to understand what is happening to that company."
The speaker illustrated with examples how generative AI could interpret and extract valuable information from an accounting table. His recommendations for prompting: make sure the chatbot understands you, give it positive reinforcement when it gets it right and personalize the query.
The challenges of AI in the tax field
During the day, in the exhibitors' area, there was also a corner for generative AI. PwC exhibited the capabilities of Harvey, with which they have been working for months. Capabilities that will be expanded with the agreement, announced during the event, with Lefebvre, which will lend its huge legal database to feed the Harvey platform exclusive to PwC Spain.
In a demonstration, Neosmart was able to see how the consulting firm uses the AI tool specialized in the legal field. They are able to summarize tax regulations and create explanatory documents for clients, as well as ask about specific documents.
Generative AI, as well as analytical or traditional AI, also featured in one of the day's round tables. Jaume Puntes, global head of tax at Cellnex Telecom, and Horacio Peña, global leader of the Payment Transfer network at PwC USA and formerly head of innovation for the company in the United States, discussed the challenges facing the tax function in the coming years.
There is a transformation in the making and Puntes says they are approaching it from two fronts. "You have to attack the data. If you don't have good raw material you are lost. We have moved away from single data and are going to truthful data," he says.
The second front is the narrative: "If you are not the one who tells the reason for this information, it will be someone else who tells it". Nowadays, the taxation of companies is being scrutinized not only by the public administration, but also by investors and even by the media.
"The global context of the new era of taxation that is coming is inevitable. There is an era of increased fees and audits, of auditing," says Peña and argues that the change is motivated by the need for postcovid recovery or the reconfiguration that companies and governments are now practicing digitally. Although one of the main factors is due to the fact that countries' debt levels are skyrocketing. There is 144% debt to GDP in the US and around 80-90% in the EU.
Peña points out that at PwC USA they work with artificial intelligence, to search for information, specific data, for translations and generation of other documents. This is the formula they have found to find a solution to the looming tax complexity: introduce technology. "We have to attack the data. Define with high precision the tax problem you want to solve. And not only use generative AI but also character recognition and other artificial intelligence systems," says Peña.