Upcoming AI Ethical Issues

The Clash of Interests, Ideology, and Control

The rapid expansion and integration of AI into most every aspect of our lives presents a new variety of unseen threats.  These threats permeate all fields raising ethical challenges in the form of discrimination, bias, privacy, opacity, autonomy, traceability/responsibility, and unjustified actions.  Nothing is immune as evidenced by theological foundations being challenged, e.g., the headline “World Economic Forum Contributor Says A.I. Could Rewrite the Bible, Create ‘Correct’ Religions.” (AI is Calling Us) https://open.substack.com/pub/robertsonc/p/ai-is-calling-us?r=v7bks&utm_campaign=post&utm_medium=web

Bias in search engines has been well documented in numerous media investigations.  The bias frequently begins with human programmer bias further skewed by data collected from an impartial internet funneled to output by the AI system.  Search engines, perceived as neutral content generators, have been shown to deliver skewed search results reinforcing racial stereotypes, gender bias, and misinformation.

In “AI the Next Religion,” a query presented to ChatGPT asked about the impacts of including Christian ethics into AI.  The ensuing response analyzed for bias and accuracy, defined Christian ethics in a human centric fashion, aligned with current cultural groupthink.  Where the program considered Christian ethics, it suggested its application as a mild influence, to “inform.”  Where the program defined religion, it made no mention of God, reinforcing our culture’s technological scientific dogma in applying AI ethics.  (AI – the Next Religion) https://open.substack.com/pub/robertsonc/p/ai-the-next-religion?r=v7bks&utm_campaign=post&utm_medium=web

Hidden within the inner workings of AI processes is the challenge of algorithmic transparency. In a prior blog, “Hidden Ways AI may be Impacting You,” https://open.substack.com/pub/robertsonc/p/hidden-ways-ai-may-be-impacting-you?r=v7bks&utm_campaign=post&utm_medium=web

examples are referenced where people were denied jobs, refused loans, put on no-fly lists, or denied benefits without knowing why, other than the decision was processed through some software.

AI privacy issues concerning data protection are just one of the increasing problems.  In March 2023 the Italian Data Protection Authority (GPDP) accused ChatGPT of unlawfully collecting users’ data and failing to prevent underage users from accessing inappropriate material.  Weeks later, OpenAI blocked ChatGPT throughout the country.  After changes were implemented, access was restored but underlying issues are leading to further challenges in other countries.

Privacy issues also extend into the ever-increasing use of surveillance.  Everyone with a smartphone camera is now a potential photojournalist.  What they miss is often recorded on street cams, satellite cameras collect everything else.  Consider an aerial view of any city, government surveillance AI can now identify an object such as a police car, then locate all other police cars.  The same technology can track you, compiling a list of places and individuals you contact.  In China, that level of surveillance is already being applied.  (AI Collective Responsibility) https://open.substack.com/pub/robertsonc/p/ai-collective-responsibility-calling?r=v7bks&utm_campaign=post&utm_medium=web

Journalism, once trusted for objective factual reporting has seen its credibility diminish with AI generated misinformation.  The problem with generative AI, used to create content, is that the program doesn’t understand the meaning of what it’s composing.  That content is finding its way into mainstream media as reported in a recent New York Times article headline, “A.I.-Generated Content Discovered on News Sites, Content Farms and Product Reviews.”

Lisa Given, an Australian Author, queried ChatGPT about her resume.  The resulting report omitted her most popular book and attributed other books that she didn’t write to her credit.

Disinformation, the deliberate use of AI to mislead, has emerged in the form of deepfake audios, videos, and pictures.  Earlier this year an assortment of fake images of Donald Trump being arrested was circulated.  Around the same time a widely publicized story emerged of an AI enabled voice cloning hoax on an Arizona mother, using her daughter’s fake voice to present a kidnapping ransom request.

Looking ahead to AI impacts in employment, the Pew Research Center estimates about 1 in 5 American workers have a job with “high exposure” to artificial intelligence.  Where technology in the past has impacted blue-collar jobs such as assembly line workers replaced by automation, future AI threatens to displace white-collar, higher paid workers.  Jobs with the lowest exposure to AI displacement are those with physical requirements that AI and robotics cannot easily replicate such as firefighters, lawn care workers, physical therapists, hair stylists, childcare workers, and dental workers.  According to the Pew Study, in 2022, workers in the most exposed jobs earned $33 per hour, on average, versus $20 in jobs with the least exposure.

In the area of creative arts, generative AI is increasingly producing content formerly delivered by human artists.  Case in point, most graphics for my blogs have been created by Adobe Firefly and Freepix. Image generators such as DALL-E2 deliver professional level graphics with little skill needed from the user.  The proliferation of easily created, low cost, creative arts threaten to disincentivize the field for creators.  Less than two years ago Github Copilot was released, it works as an AI “pair programmer” that assists in writing code.  As AI boosts the productivity of programmers, their field joins others that risk being overtaken completely.  (AI in the Marketplace) https://robertsonc.substack.com/p/ai-in-the-marketplace

Perhaps the largest sector offering AI growth potential is healthcare.  The vast amounts of information covering patient records, drug documentation, marketing data, insurance archives, and medical diagnosis/pathology, is ideally tailored to AI’s ability to sort the multitude of data points.  Potential threats from misuse of this data were reported in the New York Times article, “Warnings of a Dark Side to A.I. in Health Care.”  The article points out that AI systems will be utilized to help billing companies and insurance providers determine reimbursement payments.  The potential exists for AI to be used to alter medical information to manipulate billing in this multibillion-dollar sector.

An Invitation for Readers to Direct the Next Blog Post

The scope of this post offers a generalized overview of AI ethical issues impacting several job sectors.  A more in-depth analysis is needed to understand the more complete extent of AI challenges.  The length of these weekly blog posts limits such an analysis to one job field.  So, let’s start with one, accordingly, I’m extending an open invitation for readers to select a job field to be covered in the next posting(s).  Please send your selection to: Charrob16@comcast.net  in the Subject Line, enter “Blog Topic”.  Future posts will cover reader requests listing the requester.