Pense

“Whoever is careless with the truth in small matters cannot be trusted with important matters”― Albert Einstein

Last week, I was prepared to be shunned. This week, I'm prepared to be hated. Why? Because I agree with a person's right to cheat. Not that I agree with the decision to cheat. Indeed not. Cheating shows a severe lack of integrity and morally questionable character. But I agree with your right to decide to cheat. You have the freedom from being interfered with as you make moral decisions.

If a tool assumes the moral outcome of a decision without asking you, that is interference. No one expects their pen to stop working if they decide to write a cheatsheet that they stuff up their sleeve to take into an exam.

In fact, it is this decision that aids in the cultivation of moral fiber. Without these tests, one cannot know how you will choose to behave when the stakes are higher. As Einstein stats, the small details matter.

Then, who is responsible for stopping the act? The individual, a friend, mentor, the company or person the act is occurring to, or a third party not directly related to either. That is the AI question we will dive into this week.

Last week, I saw this post:

The post shows ChatGPT failing to respond to a user request because the tool thought they were currently taking an exam. The tool didn't know for sure. It was neither in the room nor commissioned by the school as an official tool of the exam. It was a third party tool wholly unrelated to either party.

Instead of thinking this was a win for preventing cheating, I thought 'well, I guess AI companies will be dictating morality now'. But why did I see this as an issue? Because this is a consumer application. Individual schools have the legal authority to punish cheating once caught. They have the legal authority to require students to use certain tools that will detect cheating. But this was a ChatGPT subscription not funded by the school. It was the subscription of an individual who may or may not have chosen to cheat.

You could argue that ChatGPT is a private company and as a private company has the right to refuse service to anyone. But how is this considered in practice? In US law, general public accommodations cannot refuse service to anyone. ChatGPT has made no moral or religious disclosures. So do they have the right to refuse service? How does a presumed attempt at cheating violate the corporate conscience? Which brings us to the argument on whether a tool can have First Amendment rights or what disclosures of first beliefs must companies have when a tool can act in this way. Is this a subtle move by OpenAI to advance our understanding of AI without any of the prerequisites a human would have to consider? Oh, the rabbit hole of my mind. I need some Mountain Dew. There's not enough caffeine in the world to explain the depths of riddles my mind is musing on around personal freedoms.

In the USA, we emphasize personal freedoms. We attempt to punish only those personal freedoms that would adversely affect society or another individual. Government and corporations were never intended to be moral compasses of our society. Religious organizations( Atheist or otherwise) were meant to provide moral guidance for the individual. Those individuals would then influence the development of society, not the other way around.

I'm pro restrictions, yes, on cases that affect the population en masse. I'm pro regulatory requirements applied carefully. But I'm not for the restriction of liberty. I am pro religious organizations answering questions on morality. But not a corporate entity.

This individual had every right to expect their questions to be answered by ChatGPT. Nowhere did it say 'do not use on exams or homework, we are heavily monitoring use for cheating'. That individual would then have chosen a different chatbot for their presumed attempt at cheating. ChatGPT neither knew for sure the user was cheating nor disclosed its moral leanings. If I go to a protestant, non-evangelical Christian chat app, I can expect it to provide answers along those lines. But what are ChatGPT's moral leanings? They never disclosed them. One could argue that on the outside they claim to be for general use, and general use would include cheating.

These tools are not the same as standard search engines. Their responses are shaped by an opinion many think is fact. We are unintentionally allowing our lives to be ruled by corporate propaganda that shapes its vision as utopia.

How should moral disclosures be made if we were to require them for applications that include search and respond? I don't know and I don't presume to have all the answers.

But I know I don't want this buried in the terms of service. I would want mandatory plain English disclosures during sign up and updates. I think this would help two things: (1) stop the myth that these pattern matching machines are acting on their own accord and (2) give choice to the consumer.

I'm morally against cheating. I'm also against the potential ramifications of AI companies taking moral positions without disclosure. Their responses will influence opinion throughout the world.

Shouldn't they be required to display their moral leanings publicly? If not, perhaps it's time for a resource like Ground News (not a sponsor) to rate AI companies for moral leanings so we as consumers can choose who to support. Anyone interested in a cool new research project?

So my dear reader, how do you think this conundrum should be handled?

I believe that we as the users of these tools must participate in the shaping of these tools for our society to develop into a world we wouldn't hate to live in. And a part of that is sitting within the gray and asking questions.