Through the transformations they induce, technical objects regulate behavior, they make the law. Devices based on artificial intelligence (AI) help facilitate, prohibit or change attitudes and ways of doing things. These days, there is a lot of discussion about ChatGPT, this conversation interface equipped with a powerful AI, which answers a variety of questions and also writes well-crafted texts from a few simple indications. What raise fears in several circles.
As in the world of education, where it is expected that the use of the tool will be able to deceive the methods of evaluation and, by default, empower students to use the machine to relieve themselves of the effort to write thoughtful work. But like other technological objects, ChatGPT also arouses positive reflections.
In the 1990s, with the advent of the Internet, legal scholars pointed out that the way objects ttechniques are designed and configured allows, facilitates, enables or prohibits human actions. The political choices are somehow embedded in the technologies. These establish ways of doing things that are necessary because the object is thus configured.
Sometimes these item defaults are immutable, and sometimes they allow for customization or user choice. Rather than being imposed by laws or courts, this normativity, called Lex Informatica or Lex Electronica, applies automatically, with all the advantages and problems that this entails. But these standards do not result from democratic deliberations.
Objects that regulate
If the laws of the States do not come to mark out their operating conditions, the configurations of the technical objects can supplant the decisions resulting from the democratic processes and impose themselves without our having our say.
For example, internet browsers like Chrome, Safari, or Edge, unlike a physical book or magazine, are configured to record a user’s browsing habits. There is a default rule here under which user data may be collected. This choice made by the designers of Internet browsers acts as a mandatory standard or is imposed by default, when the user has the option of modifying the configurations. For the last two decades, we have been able to observe that such possibilities of using data have had important consequences, such as the reconfiguration of the conditions for the distribution of digital advertisements. This may also have facilitated manipulations based on the targeting of individuals without their knowledge.
Through the changes it induces in the processes of production and circulation of information, AI can shake the foundations of several legal rules. Faced with the changes it induces, we will consider that state laws are or are not necessary. For example, is it acceptable to allow the use of AI-based devices to make decisions that can have serious effects on the lives of individuals? Should we not prohibit the use of such tools to respond to exams intended to verify the skills of a future professional? Technical objects can also influence the course of decision-making processes or deliberative processes. This is full of democratic issues that cannot be left to the decisions of designers alone.
An example concerning an older technology illustrates how the normative character of technology is marked out by the laws. When at the beginning of the XXe century, motor vehicles began to spread, they were accompanied by their technical configurations. By default, these configurations allowed users to drive faster than the means of transport used until then.
State laws, then designed to regulate the circulation of horse-drawn vehicles, found themselves supplanted by the new possibilities arising from vehicle configurations which, by default, could travel at higher speeds. To respond to the consequences generated by these possibilities provided by these technical objects, the laws have been updated in order to regulate the conditions of use of vehicles equipped with engines.
Like state laws, the configurations of technical objects are the result of political decisions. By recognizing the normative character of technical configurations, we give ourselves the possibility of debating the legitimacy of these standards implanted in the architecture of objects and which impose themselves without discussion.
A technology like ChatGPT contributes to modify the ways of considering several human activities. This may or may not coincide with what is held to be desirable or acceptable. When we recognize that technical objects are accompanied by their normativity, we can question the adequacy or harmful character of the laws that their configurations impose on everyone. It is normal and legitimate to subordinate the regulation imposed by technical objects to laws adopted democratically. Unless we resolve to leave only technologists the ability to impose their choices on us.