Chapter 2
AI in real life

Despite the risks, ending AI is not being considered; Get to Know its practical uses

The benefits people may get from AI outshine ethical issues

Raphael Hernandes

Despite the doubts about the impact of AI (artificial intelligence) on human life and its risks, abolishing the technology has not been considered.

"Never in history has a technology stopped being used for ethical reasons," said Renato Rocha Souza, a professor at FGV and a researcher at the Austrian Academy of Sciences.

The reason for the support is that AI can bring benefits in different areas. One of which is medicine.

The technology can act as an aid to doctors in detecting diseases when suggesting diagnoses or treatments. The decision, in the end, must remain in the hands of the professional.

Artificial intelligence can find patterns in millions and millions of mammograms, for example. In this way, it can help to identify a tumor in an imaging exam.

At Hospital 9 de Julho, in São Paulo, the AI ​​monitors images from cameras installed in the rooms and detects movements considered to be risky, such as the lowered bed rail or a patient lying very close to the bedside. When this happens, an alert is issued for the nursing staff to intervene to prevent a fall.

University of Oxford's AIxSDGs project is surveying real cases of AI uses for humanity, and the list will launch in September. Researchers gather initiatives around the planet that use artificial intelligence to help fulfill the UN's Sustainable Development Goals.

In one month, more than 80 projects were cataloged on the platform. They include solutions like the Indian ConserWater, focused on reducing water waste in plantations, and the British Zooniverse, which analyzes satellite images to find slave work areas.

In the industry, one of the main applications for AI is to analyze data coming from sensors spread throughout the production steps.

HI, I AM THE ROBOT

Another widespread use of AI is content recommendation. This is the feature that allows Netflix and Spotify to suggest movies and music based on the customer's consumption history and comparing their profile to others of similar tastes.

Speech recognition mechanisms present in assistants on smartphones also use artificial intelligence.

It is not magic that makes what is said appear in the form of text written on the screen of the cell phone, but NLP (acronym for "natural language processing"). This is also one of the parts of the arsenal used for one of the most popular AI applications today: chatbots, or robots for service.

Anyone who sees these mechanisms working may not think of all the complex architecture under the hood. In addition to processing and interpreting what is said, they depend on connections with different sources of information to be able to generate the texts of the responses.

In a chatbot linked to a financial institution, for example, when a person asks "how much money do I have", he needs to understand that the intention is to consult the account balance, even if this has not been said explicitly.

From there, the system communicates with the bank's data to discover the information and be able to turn it into a response.

This dynamic improves over time, in a learning process. "Our entire experience doesn't start 100% on the first day," said Leonardo González, director of data and AI at IBM Latin America.

HUMAN REPLACEMENT

One of the concerns artificial intelligence sparks is the replacement of human labor by robots. Studies show that some professions will disappear, a common phenomenon with the advent of new technologies since the beginning of time.

A research study published in January 2019 by the Machine Learning Laboratory in Finance and Organizations at UnB (University of Brasília) pointed out that robots threaten 54% of jobs in Brazil. Technology threatens the jobs of chemical engineers, warehouse loaders, and even volleyball referees.

In some cases, however, new technology leads to new functions. At Vivo, people who used to work in customer service now work on the Aura chatbot.

Luiz Felipe left the call center to become a robot teacher. He underwent an internal selection and subsequently underwent two-week training to take on the new role.

"We see if Aura will understand the client's speech. For example, in the South, they say 'boleta' instead of 'fatura,'" said Felipe, citing a situation in which human intervention is necessary to improve the service –both words are used in Portuguese to refer to invoices.

Translated by Kiratiana Freelon

Read the article in the original language