Clara Grima, mathematician: “We are going to do what machines cannot”

Recent advances in technology and artificial intelligence have brought into popular conversation a term that until recently was reserved for mathematical contexts, a term that most people only remembered, if at all, from a few math classes in high school.

Algorithms govern our virtual lives, and this has earned them a bad reputation, linked to biased, dangerous or harmful use. Shedding this reputation is the aim of the latest book byClara Grima, PhD in Mathematics and Professor of Applied Mathematics at the University of Seville, entitled Con algoritmos y a lo loco (With Algorithms, It’s Crazy).

Subtitled “Because algorithms aren’t as bad as they seem,” the book reviews the history of these ordered sets of instructions that underpin much of our daily routine, from checking accounts in a banking app to navigating to an unfamiliar destination. “The word ‘algorithm’ wasn’t familiar to the general public until the advent of computer algorithms, when it is the first thing you learn in math class,” explains Grima, who combines her teaching and research work with science communication. Algorithms are operations as basic as addition and subtraction; however, the term’s use has been popularized by its misuse in digital environments, from its application in social media to music recommendations. But in themselves “algorithms are neither good nor bad,” argues Grima. Like almost everything in technology, it depends on how they are used.

In fact, Grima does not hesitate to label some algorithms “works of art.” This is the case with Fast Fourier Transform, or FFT, which has countless applications in everyday life: from image and audio processing to the optimization of wireless networks or its use in the fields of scientific and medical measurement and analysis. Or Google’s algorithm, which “should be in museums, because it is a marvel that with very simple mathematics, it does what it does.” There are also other algorithms that allow the calculation of shortest paths, which we always think of in terms of logistics and business, but they are also used in the distribution of humanitarian aid.” Or those on which cryptographic systems are based that allow information to travel securely when accessing a banking app. These are a few of several examples that Grima champions “because they are beautiful from a mathematical point of view and because they are very useful for many things in our lives.”

The need for mathematical talent

One of the biggest problems in today’s tech ecosystem is attracting tech talent, which largely comes through training in mathematics. The negative perception of algorithms also plays a role here, Grima reflects. “Children enter school, and even before studying math, many already have math anxiety,” she explains. “If we further increase their fear by portraying it as malevolent, it only intensifies that anxiety.”

The problem is no longer limited to that early stage of education; it persists into high school and affects university aspirations. “The job market, with all the artificial intelligence and the revolution we’re experiencing, needs many people who know math. So, this math anxiety can ultimately lead to our children being excluded from the workforce,” she continues. That is all-the-more reason to advocate for a positive view of algorithms, “besides appreciating their beauty, their usefulness, and the mathematical language and concepts that always accompany them.” Also to link this discipline to everyday life, because “the basis of all science and technology is always mathematics.”

In this field, as in other STEM careers, a significant gender gap is evident, one that has developed as mathematics has become a prestigious and career-oriented field of study. While in the 1985-86 academic year the number of women studying mathematics at public universities in Spain exceeded that of men (4,414 versus 4,295), this trend reversed from the 2005-06 academic year onward, reaching 5,020 women compared to 8,707 men in 2022-23, according to data from the Spanish government cited by the Royal Spanish Mathematical Society. Grima confirms this trend, which she describes as “a little demoralizing.”

“If there aren’t more women in computer science, and on top of that, the number of women in mathematics is declining, these algorithms they have to supervise are going to be very biased.” It’s not a matter of men performing worse, she argues, “it’s that they have a different sensitivity.” She illustrates this with her own experience, explaining how, since becoming a mother and having to push a stroller, she became aware of the problems with building access ramps.

“Everyone has a different environment and needs, and it’s crucial that all perspectives are considered in the design of algorithms.” Hence her concern about current developments: “The fear that these algorithms, which will control or decide many public matters, will be biased if the groups of people working on them aren’t diverse. And they aren’t.”

Algorithms to control algorithms

Once it’s clarified that algorithms are neither inherently good nor bad, and their positive uses have been acknowledged, we must also consider the scenario of real-world misuse, for which Grima also has a solution: more algorithms.

“We need algorithms that control algorithms,” she explains, though reluctant to use the word “control,” correcting it to “supervise.”

“This supervision must be carried out by other algorithms,” she emphasizes, although its development requires a multifaceted approach. “There’s a problem that isn’t about mathematics or computer science, but about ethics and philosophy. It’s about reaching an agreement, a global ethic.”

Grima speaks of a process similar to the development of press legislation, in which minimum standards are established for what to do and what not to do—even though certain media outlets then disregard them. It’s about establishing this global ethics “to define and design algorithms” and to limit the risks involved in their implementation.

In this context, Grima champions the role of the humanities, of disciplines such as philosophy, linguistics, and history. “Humanity has advanced so much in science and technology that we now have a world which is not what it used to be, and it needs us to sit down and think about how we are going to manage these tools,” she argues, emphasizing that “human beings won the evolutionary battle because they were empathetic, because they formed groups, and because they cared for one another,” which makes it necessary to rethink this role.

“Since machines are going to do many things for us, let’s do what they can’t: think and sit down and talk like humans.”Agentic AI: What now, what next? – ComputerworldRead More