Artificial intelligence in 2021: innovation and challenges for businesses

Covid-19 has changed the world as we know it and has affected all industries. One of those is artificial intelligence (AI), which has increased in relevance and grown exponentially during recent months. “The pandemic has forced companies to take a leap of faith into digitisation that some reports had forecast for three years away, by 2023”, explains Juan Murillo, member of the Data Strategy team at BBVA.

In September 2019, consulting firm IDC had forecast that in 2023 worldwide spending on artificial intelligence systems would reach around $98 billion (over € 81.9 billion), 2.5 times more than expected for two years ago, before the world changed.

A survey published by the McKinsey consultancy in November 2020 concluded that companies were using AI “as a tool to generate value” during the pandemic. In fact, of the almost 2400 respondents across all sectors and regions, half explained that they used it for “at least one function” of their operations.

Nevertheless there are still some challenges, such as achieving general AI that doesn’t just solve specific problems, as happens now. That is why some are predicting the arrival of super artificial intelligence, minds made to be similar to those of humans. Assisted by Murillo and the other experts who participated in the ‘BBVA Open Talks Global: Algorithms for a new era’ event, we present the state of AI today and how it can generate innovation in companies of all sizes.

AI helping to manage a pandemic

Coming back to coronavirus, Edwin Villanueva, a researcher for the Artificial Intelligence Group at the Pontifical Catholic University of Peru explains that “the current challenges around AI are to do with their application: how can we get AI to help us with this pandemic, optimising the application of vaccines and logistics, and making better diagnoses?”.

Sofía Trejo, head of educational and outreach projects for the Artificial Intelligence Alliance of Mexico’s National Council of Science and Technology, indicates that since 2020 global private investment in the pharmaceutical and health sectors has increased: “This type of AI is being used for cancer treatments, new medications and molecular synthesis. Recently, AI modelling was used to determine protein structures in 3D, something that was previously done with expensive experiments”.

According to the McKinsey study mentioned earlier, representatives of healthcare services and pharmaceutical and medical companies, together with those from the automotive and assembly industries, were those who claimed to have most increased investment in AI.

Murillo also sees in AI a challenge generated by the pandemic itself in training models when the data has changed in such disruptive months: “You have to refine this type of analysis because, if they are trained with the data from a very abnormal year like 2020, they could generate incorrect conclusions. Predictive analyses need to eke out anomalies and know how the economic context may evolve in the near future”.

Data and customer digitisation

The lockdowns prevented many people from accessing bank branches to carry out transactions, but they found out that these could be done digitally. This gave banks the opportunity to obtain large amounts of data to improve their business.

“We’re in a virtuous circle“, explains BBVA’s Murillo, about one of the business and banking trends we will see in 2021, adding: “Once customers have seen that for the majority of their operations they can easily use digital channels, their digitisation is consolidated, meaning that the volume of data generated is increasing”. In this way AI itself will improve, because it feeds on that data. “Since we’re increasingly digitised as citizens and as customers, we have more raw material to train new models, which can translate into them being more accurate”, notes Murillo.

Among the banking activities that benefit from AI, Murillo cites financial health advice based on previous forecasts, risk analysis to improve the management of investment portfolios, and even help against money laundering.

Stock exchanges can also take advantage of advances in AI. Villanueva, from the Pontifical Catholic University in Peru, explains that “business algorithms were wrong [due to the unforeseen economic context with COVID-19], but those that have the ability to learn quickly are generating better results”.

AI helping companies to innovate

After the pandemic has forced means of remote contact to improve, Trejo, from the Artificial Intelligence Alliance of Mexico’s National Council of Science and Technology, believes that we will see new developments in customer service AI. This is evidenced by chatbots that exist to solve queries and free up space for humans to focus on less trivial tasks.

AI also allows companies to innovate in other internal processes. According to Villanueva, “companies (for example, manufacturing industries) have numerous probblems, particularly in resource optimisation. AI can help them operate, organise workflows, perfect the use of machinery and streamline logistics”.

In Murillo’s view, it is not only large corporations, who we assume have huge amounts of data to manage, that can train their algorithms. For small and medium-sized enterprises, there are European Digital Innovation Hubs, which have come about as a result of European strategies in the areas of digitisation, data and artificial intelligence and that “promote the adoption of artificial intelligence in SMEs”.

Innovation without losing sight of ethics

All of these business innovations cannot be developed without ethical considerations. As criticism of algorithm bias in policing, justice and access to social services increases, demands for fair and independent AI audits are on the rise. Trejo highlights international initiatives such as the UNESCO recommendations and the Global Partnership on Artificial Intelligence (GPAI), a group to which Spain, Mexico, the EU, the United States, South Korea and Japan already belong, among other countries. The GPAI works for international alliances to ensure responsible development of AI.

The Mexican makes a wish for 2021: “I would like governments to bebgin regulating this technology this year, based on these recommendations. I hope that external, independent auditing will begin in some form or another.”

Villanueva says that we’re experiencing an “industrial revolution” that may steer humans “towards activities that machines cannot automate, such as political decisions, art and creativity” and that in the medium term we’ll see “AI that is stronger and more sophisticated“. So it is possible that the technology will keep advancing to achieve the long-awaited general AI that currently appears unattainable.