The Last Stronghold: Sacred Professions Resistant to AI

The rapid and ongoing integration of AI (Artificial Intelligence) across industries has raised the possibility of a future where AI could potentially replace a substantial portion of jobs. A report by Goldman Sachs suggests that AI has the potential to replace the equivalent of 300 million full-time jobs. This finding, highlighted by the Goldman Sachs report, underscores the magnitude of potential job displacement that AI could bring about.
To make it very clear, this transformation is not projected to anticipate in the distance future such as 20 or 30 years from now; rather, it is expected to transpire in just a few short years, as global powers compete in a race to achieve a breakthrough in intelligence that will exert dominance on a worldwide scale.
Another crucial point to emphasize is that this is not mere speculation; it is a certainty that AI will soon replace a significant number of jobs.
The impact of AI on work tasks is also a significant consideration. According to the Goldman Sachs report, AI could replace up to a quarter of work tasks in the US and Europe, highlighting the breadth of sectors that could be influenced by automation.
Consider the music industry as an example: AI has the potential to make a substantial impact by automating various tasks, potentially reducing the reliance on manual labor. Here's an illustration:
AI-powered algorithms and machine learning models can analyze vast amounts of music data, including popular songs and trends. AI can generate original music compositions, melodies, and lyrics based on this analysis. This can reduce the need for human composers and lyricists in creating music, especially for background music, jingles, or repetitive commercial tunes.
AI can also personalize music composition for individual users or specific demographics, further reducing the requirement for human composers to create customized music.
Additionally, AI-driven tools can automatically generate sheet music or transcriptions from audio recordings, simplifying the work of music composers.

The Influence of AI on the Jewish Workforce
Interestingly, perhaps one of the few work fields worldwide that remains entirely impervious to automation is the domain of Jewish religious professions. Consider, for instance, the intricate tasks like inscribing sacred texts such as Sefer Torah, Tefillin, and Mezuzot. These items must be meticulously written by a Jewish sofer and cannot be substituted by any other means. Another example pertains to the intricate processes involved in slaughtering animals, which may only be performed by a shochet possessing the requisite qualifications and intention.
Similarly, the same applies to the preparation of Tzitzit, which necessitates the intention to fulfill the mitzvah of tzitzit. Furthermore, the craft of making Matza (shmura) for Pesach serves as another prominent example that must be carried out by a Jew with very specific thoughts in mind during the preparation process. The absence of these intentions and thoughts renders them invalid.
There are numerous additional examples, including the blowing of the shofar and the practice of brit milah, among various others.
The resilience of these fields can be attributed to two key factors. Firstly, those engaged in these professions must be observant Jews who adhere to Torah law. Secondly, they must infuse their work with intention, particularly the intention required to prepare the mitzvah object "lishma," meaning solely for the purpose of the mitzvah. This level of dedication is a quality that AI lacks. While artificial intelligence may display remarkable intelligence, it doesn't possess the spiritual commitment and intention that are integral to these sacred practices.

Can machines prepare the Mitzvah items?
When machines were first invented, rabbinic authorities (poskim) grappled with their usage in the preparation of mitzvah items, including the ones mentioned earlier. Numerous poskim maintained that machines are not valid for this purpose due to their inability to meet the two conditions we previously discussed; furthermore, they lack the capacity to prepare them "lishma" since this requires intention, which AI lacks. Some poskim, however, argued that it's sufficient for the person to have the necessary intentions before initiating the machine, and when the machine begins preparing the items, it's considered as if it's done "lishma."
The poskim derived this concept from the teachings of the Rosh, who maintained that it is permissible to instruct a gentile to begin a task with the intent of fulfilling a mitzvah (lishma). This perspective contrasts with the opposing view of the Rambam, who asserted that any involvement of a gentile in the work would render the item unfit for a mitzvah.
In accordance with this understanding, just as one may direct a gentile to perform an action lishma, similarly, one may also guide a machine's operation by specifying that it should perform its task with the same lishma intent.
However, the Chazon Ish clarifies that this principle applies only when the person performing the task visibly appears to be doing it on behalf of the sender. In typical cases where there is no suspicion of a change of heart, the task is still considered lishma because it consistently appears as if it is being performed for the sender. For example, in the case of writing a get, where the husband is likely to change his mind and reconcile with his wife, it does not appear that the gentile is writing it for the husband's benefit.
Is AI less proficient than machines in the torah aspect?
Understanding the novelty of AI and how it differs from traditional machines is crucial for our discussion.
AI differs from traditional machines in terms of its ability to learn, adapt, and make decisions autonomously. Traditional machines operate based on pre-programmed instructions and repeat those instructions without deviation. They lack the capacity to learn from new data or experiences or to make independent decisions beyond their programmed parameters.
In contrast, AI systems, particularly those powered by machine learning and deep learning algorithms, can analyze data, recognize patterns, and make predictions or decisions based on that data. AI can adapt and improve its performance over time as it encounters new information, allowing it to operate in a more flexible and dynamic manner.

Now, let's apply this perspective to AI:
On one hand, it can be argued that AI systems are tools or extensions created and controlled by humans. When a person employs AI to execute tasks or actions aligned with a mitzvah, they are essentially guiding the AI to continue the task with the original lishma intent. In this sense, AI could be considered an extension of the person's will and intent.

On the other hand, since AI represents a groundbreaking departure from conventional machines, it possesses the ability to learn, adapt, and make decisions autonomously, often resembling human cognitive processes. An opposing viewpoint suggests that AI's autonomous decision-making capacity sets it apart. It may not appear to perform tasks on our behalf in the same manner as traditional machines. According to the teachings of the Chazon Ish, if it doesn't seem like the AI is acting for us, it may be considered invalid. In contrast to a non-Jew preparing on our behalf, where we have no reason to suspect a change from the original intention since they have no motive to do so, AI continually learns and enhances itself. Consequently, one could argue that it doesn't appear as though it's performing our task.

This presents a distinctive challenge when evaluating the validity of AI in the context of mitzvot. Even if someone may argue against the aforementioned assumption and assert that AI is equivalent to machines, there remains at least a level of doubt. Engaging in mitzvot with doubt is generally not permitted. Additionally, there is a disagreement (machloket) regarding the validity of mitzvot performed with machines, and most poskim maintain that they are not considered kosher.