AI’s legal framework will be a consumer protection resource – 03/21/2023 – Maria Inês Dolci

Artificial intelligence (AI) has everything to do with the CDC (Consumer Protection Code). Therefore, a legal framework for Artificial Intelligence will certainly have to relate to the CDC, the Marco Civil da Internet and the LGPD (General Law for the Protection of Personal Data). I do not rule out that, at some point, we will have to include articles in the CDC so that this link is clearer and more productive.

The first connection occurs in the right to information, one of the highlights of the CDC. The consumer has the legal right to be informed if AI was used in the provision of services, in marketing, in mapping purchase intentions and in the manufacture of a product.

It is also our right to know whether, in consumer situations, we are interacting with AI. And there must be complete clarity about, for example, the reasons why we were denied financing for a particular purchase. What criteria are used? How does an algorithm define higher or lower risk of default?

There are situations in which we have the impression that our thoughts are being read. In fact, we are tracked all the time. A simple query about a product causes specific advertisements to appear on each access to a social network.

Other aspects to be considered are the functioning of products and services. Who is responsible for an accident due to technical problems in an autonomous car? I advance that, independent or conventional, the responsibility will be solidary, that is, of who manufactured and sold.

And what about robots that care for the elderly? And if they, for some reason, make a mistake in administering a medicine, harming the health of the person being cared for? Again, the responsibility will be divided between manufacturer and seller. This would also occur in a surgery using robotics, which would cause complications and even death, involving the doctor, hospital and manufacturer.

It may also be necessary to better standardize virtual customer service. The attendant robot can only answer trivial questions, previously programmed. In the face of more complex cases, it should be immediately transferred to a human attendant. I also defend that the consumer can choose to be served or not by a robot.

Inadequate service by an automated application will be the responsibility of whoever developed the virtual attendant and the company that purchased the app.

In December, the President of the Senate, Rodrigo Pacheco (PSD-MG), was handed over a proposal for a substitute bill to regulate the development and use of AI in Brazil. The text contains several governance obligations for vendors and operators of AI systems. According to Valor Econômico newspaper, this study was prepared by a commission of jurists appointed by the Senate.

There is a lot of work ahead for the definition of modern legislation, with a focus on citizens’ rights. We will still be dealing with this subject a lot.

PRESENT LINK: Did you like this text? Subscriber can release five free hits of any link per day. Just click the blue F below.