Galileo Era Blindspots
In the era of Galileo Galilei, around the early 17th century, if AI were available, the trial that sentenced him for heresy might have unfolded in far different ways, depending on the available data. Galileo’s radical proposition, arguing in favor of a heliocentric universe contrary to the geocentric worldview held by the Catholic Church, was a pivotal point in the history of science.
Suppose the AI of that time was trained predominantly on religious texts, philosophical works echoing geocentric beliefs, and scarce scientific literature endorsing the heliocentric model. In this scenario, the AI might have mirrored the predominant sentiment of society and the Church. Its lack of access to an extensive body of scientific work supporting Galileo’s views could have led the AI to assert that Galileo’s claims were baseless or radical, reinforcing the Church’s stance. This ‘blindspot’ might have further solidified the opposition to Galileo, potentially leading to an even stricter sentence.
On the other hand, if the AI had been exposed to a balanced data set, including a wider range of scientific observations and arguments supporting the heliocentric model, it could have served as an impartial arbitrator. It could have articulated Galileo’s ideas in a manner more palatable to the Church, focusing on the evidence and the spirit of inquiry rather than the perceived heresy.
In this scenario, the AI could have bridged the gap between dogmatic belief and empirical evidence, potentially altering the course of Galileo’s trial.
Gaping Holes in Our AI Legal Analysis
Artificial Intelligence has increasingly been harnessed to provide legal advice, from understanding contractual clauses to interpreting court decisions. However, a critical limitation constrains the accuracy and relevance of these AI systems: the inability to access all US laws, ordinances, and judicial decisions.
As it stands, the AI knowledge base comprises a fraction of the full breadth of American laws. Even though federal and state laws are mostly available online, a substantial portion of local ordinances, rules, and judicial decisions – particularly from smaller jurisdictions – remain offline. This selective accessibility to legal data creates “blind spots” in AI’s legal knowledge, which may skew the results they produce, leading to potentially incorrect or incomplete advice.
These blind spots manifest in multiple ways. An AI legal assistant may provide advice based on federal or state law while missing a critical city ordinance that could materially affect the advice given. Similarly, an AI trying to predict the outcome of a lawsuit may overlook a relevant judicial decision from a smaller jurisdiction because it’s not in the training data. These discrepancies might not only result in less accurate legal advice but also breed mistrust among users who discover these oversights.
Furthermore, this lack of comprehensive data perpetuates a form of ‘legal centralism.’ The AI’s knowledge is skewed toward laws and decisions from larger, more technologically advanced jurisdictions that have the resources to digitize and publish their legal documents. This bias may reinforce a perception that legal advice or decisions from these jurisdictions are more valid or authoritative, sidelining the legal norms and precedents from less represented regions.