Artificial intelligence in Brazilian health and supplementary health services

The first article of our series, Artificial Intelligence: Impacts on Healthcare and Agribusiness,  addressed the use of Artificial Intelligence (AI) in software and medical devices.

This second article addresses aspects of the use of AI in health services in Brazil. It looks at how the Brazilian Supplementary Health Agency (ANS) has been leading a digital transformation within the country’s health sector, as well as important updates concerning the regulation of AI in Brazil, including a substitute bill for an AI act that was presented in November 2023.

AI and health services

Given the lack of specific regulations in Brazil, professional councils have become an important source of information for health professionals in regard to applying AI to health services. In response to an increased supply of AI-based products and services, these councils have focused on publishing guidelines or making commitments to address the issue.

One example is the Brasília Charter (Carta de Brasília), a document that establishes bioethical commitments for Brazilian health professionals in response to technological advances in the field of health and medicine. The Brasília Charter was published after important events held by the Federal and Regional Councils of Medicine. It emphasizes the importance of incorporating the study of AI into academic curricula at all levels of training to provoke critical analysis of AI tools and, consequently, ensure that implementing this technology does not weaken the doctor-patient relationship and the quality and safety of patient care.

Other professions linked to the health sector are also analyzing this question, discussing whether they should (and how they can) use AI tools. These tools could be applied across a range of activities, from providing functional support for medical practices to effectively contributing to the core activities of healthcare services (including assistance with surgical procedures, diagnoses, testing and determining appropriate treatment).

In another example of how AI occupies an increasingly prominent space within the sector’s daily activities, in 2023, Brazil’s Regional Pharmacy and Nutritionist Councils (CRF/CRN) also debated the benefits and limits of applying AI in their respective areas. Furthermore, the Higher Education Improvement Office (CAPES) – a body linked to the Ministry of Education – released a scientific study that applies AI systems to detect tooth decay via photoluminescence in digital image processing.

Considering how AI is attracting interest worldwide and is appearing more and more throughout different service chains, the lack of specific legislation is no reason for postponing or relativizing (bio)ethical scrutiny and questioning the creation of AI tools for the health sector.

From the outset, the development of any AI solution should consider principles such as the protection of human dignity and of personal data, non-discrimination, and the establishment of informed consent and transparency in relation to AI system processes and decisions. In addition to the precautions developers should observe, companies that contract, implement, and offer these solutions to end users should first evaluate the possible adverse impacts on both individuals and society as a whole.

This debate is constantly evolving and escalating, and rules are expected to be set in the future. Irrespective of this, parties that work with AI systems (whether as system developers or their contractors) are recommended to already take necessary precautions to include the specifications of the contracted solution and explicit provisions about liability for damages that may stem from their use in their contractual instruments.

AI and supplementary health

In the context of supplementary health, the ANS is committed to digitally transforming its services. This can be seen in the Digital Transformation Plan, which is part of Brazil’s Digital Government Strategy. The plan specifically aims to expand the availability of digital services, as well as increase user satisfaction with digital services, simplify bureaucracy and reduce red tape, reduce average waiting times for ANS services, promote integrated use of data science, and implement technological solutions that optimize workflows.

In 2023, the ANS signed contracts with three startups to develop the following AI projects as part of the AI Solutions for the Public Sector initiative:

• A tool for analyzing online complaints that seeks to speed up response time for health insurance beneficiaries who submit requests to the ANS to guarantee their right to access healthcare;
• A digital assistant for modifying regulated companies’ registration data, seeking to streamline the ANS’ performance in regard to the economic and financial supervision and regulation of health insurance operators; and
• The use of AI in the public health system’s (SUS) reimbursement procedures.

In addition to the application of AI in ANS’s administrative activities, health insurance operators and other entities in the supplementary health sector have also been using AI in their operations. Such applications include efforts to boost integration between the offices of health insurance operators and their partners to improve patient guidance and monitoring, optimize medical procedure authorization processes, create a chatbot for quick and free access to information about the supplementary health sector, and tackle fraud in supplementary healthcare.

Regulation of AI in Brazil

Debates on an official Brazilian AI Act are currently underway in Congress, with the goal of defining guidelines, principles and rules for the responsible development and use of AI systems.

In the wake of discussions held by the Committee of Jurists, established by the Federal Senate, to address the topic last year, Bill No. 2,338/2022 – largely inspired by the European Union’s AI Act – was proposed by Senator Rodrigo Pacheco.

However, on November 28, 2023, an amendment was introduced to replace the bill with a completely new version. The main changes included removing a chapter on the rights of people subject to AI systems and removing provisions for a specific methodology to assess the risks associated with these systems. Moreover, while the original text introduced a list of high-risk AI systems (based on the purpose of their use and the sector they would be applied in), the new version provides parameters for assessing the likelihood and the negative impacts of the AI system, as well as parameters for calculating and classifying the risk (as high, medium or low) of using the system.

As per the new proposal, contributing factors to assessing the likelihood include system capacity analysis, the types and quality of data, system quality, cyber vulnerability, and ease of use and access. Meanwhile, contributing factors to assessing an AI system’s negative impacts include how it affects human beings, national sovereignty and security, as well as its financial, material and environmental impacts. The substitute bill exemplifies that certain systems that are critical to society – such as health – can be assessed as having a medium-high impact.

At present, the lack of a single, unified initiative for regulating AI makes it difficult to establish a clear prognosis on the final provisions and timeframe of enacting specific AI legislation. However, due to its relevance to the health sector, it is clear that this matter is of high interest to Brazilian legislators and is expected to be the focus of further discussions in the near future.

The next article of this series will explore the legal aspects of using AI systems in Brazil’s agricultural industry, as well as the impact of the proposed regulations, challenges, and opportunities in this sector.

For more information on these topics, please contact Mattos Filho’s Technology and Life Sciences & Healthcare practice areas.