Progress: Confession to an AI Priest
Terry H. Schwadron
May 3, 2024
We’re already seeing headlines that show Artificial Intelligence starting to have real effects on how we pursue our jobs, tasks, and cultural habits.
The early signs are a mix of wonderful and, well, strange.
There are the predictable tech advances in making routine matters more mechanical, along with the expected warnings about too much dependency on machine-produced answers. That is leading to worries about students using machines to write their assigned essays or the oddity of AI-driven human resource machine reviews of AI-produced job resumes.
Of course, even before AI, there was already a longstanding tradition of moving towards automation in designated jobs. In California, which just raised minimum wage for fast-food workers, we also saw the formal opening of the first all-machine produced burger shop in Pasadena. AI provides machines that not only automate, but learn, making it a contender for other kinds of jobs.
AI already is proving useful in medical applications including those accelerating or making more precise various uses of radiology tests and guiding surgical tools, in research towards earlier detection of some diseases, in treatment design for gene editing and other such areas.
We’re also getting an early dose of AI-doctored political advertising with collages of voice and video samples that are good enough to pass an unsuspecting eyeball — undoubtably further eroding any trust in those seeking to run the very governments that seem not to know how to bring regulation to excesses in artificial intelligence development.
According to CNBC, which was relying on data from ResumeBuilder, more than one-third business leaders nationally say AI replaced workers last year, with more each month. The job market is adjusting as more work is declared machine-doable, even though someone must tell the machines what to do and program them. Most are manufacturing jobs, where some assembly line jobs can be replicated with programmable procedures.
Politico mentioned several examples of AI’s rawness in finding odd conclusions from data, citing lawsuits about AI background check companies have been sued for making mistakes in job reviews. Some resume screening tools have been found to perpetuate bias, another identified as two factors of future job performance to be having played lacrosse and being named Jared. Another assessment provided high scores in English proficiency even when questions were answered exclusively in German.
We are at such an early stage of artificial intelligence technology that we have no way really of measuring its impact. But we ought to be on alert for errant results. Still, AI has proved itself already to be the lure of international states jockeying for financial and technology supremacy, since it is recognized as a critical economic foundation in the years ahead.
AI Caring
Recent reports have focused on services where we expect a human face, like counseling and mental health services. Stories on 60 Minutes and in major news outlets have detailed early efforts to establish online chats with AI-fed machinery to offer person-less contact for questions and answers about a range of personal issues for which there may be more standard responses.
Obviously, as psychology appointments are becoming more difficult and expensive to arranger, AI chatbots can offer some advice for issues it has been fed to understand. Professionals say that under certain circumstances, offering some standard responses to the most common mental health complaints can be helpful, but the coding almost always tells potential patients to consider seeking personal intervention with a human, who can better assess and respond with non-machine caring and empathy, for example.
It is interesting to note that ethical issues aside, these AI approaches also have engendered discussion about whether they err in failing to recognize cultural, racial, or gender-oriented aspects of depression or other emotional disruptions. Of course, reading books about depression or even searching online resources have always helped to fill this gap.
This week, there was yet another oddball front on the AI-human caring-services front — a small backlash to a virtual AI priest called “Father Justin” who was appearing online to offer absolution from reported sins. Catholic advocacy group Catholic Answers released the AI cleric but then “de-frocked” him when users complained — that is, the advice continues, but without a patina of online identity as an actual priest.
Apparently, the app was offered to help people explore Catholic faith. It’s advice was pretty conservative, even outdated, said too many users, and the priest presented himself as a real clergyman raised in Italy and taking online confessions and offering some a sacrament.
Chris Costello, IT director at Catholic Answers, said “We believe that the presentation of Father Justin honors real-life priests and the role they play in people’s lives, yet we are confident that our users will not mistake the AI for a human being.”
I must be missing something: Why not just go to church? Where’s the need that is better served by confession from home?
A Humbler Approach
We’re caught between imagining a robotic world in which machines take over and one in which we can’t believe we’re not allowing sufficient automation tools to eliminate the humdrum or to speed advances in business, health, education, and other processes.
The companies making artificial intelligence machines, self-learning engines that gobble up whatever is fed to generate their own newly created language, freely admit that the coding is running ahead of their own understanding of the results and how to control them.
What we once envisioned in wars being fought without soldiers increasingly depicts the reality of drones and over-the-horizon weaponry. What we saw as rigor in medical research is being supplanted by speedier machine reviews of layered health testing. What we have seen in processes from inventory controls and re-ordering to financial analyses and productivity reviews is changing by the day.
It does not take huge sci-fi imagination to think that driverless cars on powered highways could offer safer transportation, or that machine valets and household robots could take over repeated tasks. But there always is a price. Our turn to become a cashless society means that our children grow up without knowing — or thinking that they need to know — how to make change from a purchase, for example, and learning how to approach art, music, dance, or writing should just be a matter of pushing an automated button, not the result of creative work.
Our lawmakers don’t have a clue what to do, or even what to regulate. The Senate has set up a year-long review of the issues, just to get a handle on what is regulatable. And even then, the onus is falling on the excesses, like fraudulent use of political voices and images.
The idea of promoting AI clergy is so wacky that it just doesn’t fall easily into any category that would merit lawmakers’ time.
The least we could be doing before we launch these new AI programs to address human suffering or confusion is to consider why we have such services in the first place. Then maybe we can decide whether machine replication and replacement are even appropriate.
##