AI and real breasts

In the 1991 movie, LA Story, Steve Martin’s character is in bed with the character portrayed by Sarah Jessica Parker. Martin says, “Your breasts feel weird.“ To which Parker casually replies, “That’s because they’re real.“ We’ve reached that point with commercial writing and generative artificial intelligence. Some clients are so paranoid that contractors might be using AI, that they’re calling out human-written content because it sticks out from the gray, banal, cliché-filled drivel that they have become used to in the corporate world.

Image from ChatGPT

This demonstrates a couple of things. One, is that AI tools are starting to become ubiquitous and the other that the corporate standard is not “the best available.“ It’s “good enough“ or worse, “what we’re used to.“ That's great news for AI companies. The statistical middle is where large language models excel. They’ve been trained on decades of average writing and corporate speak. They’ve literally been training all of their silicon “lives” for this moment.

It was much the same in the 1980s when word processors became widely available. There was lots of mediocre writing because most people are mediocre writers. In the 1990s, it was the same when consumer page-layout software debuted, because most people only think that they are graphic artists (think Comic Sans.) Ironically, the above is the literal cause of the banality of AI tools–all of that mediocrity was used to train them.

Some writers have taken to including disclaimers that their material was human written. Of course, the next addition to AI prompts will be to routinely include such a disclaimer even in AI-generated content. This struggle can only end with a general acceptance of AI-based writing tools.

In these dialogues about AI involvement in social media posts, many people overestimate their ability to perceive humanity. Generative AI will certainly continue to improve its ability to effectively simulate human origin. If readers can’t accurately discern AI sourcing, then that value will be based purely on content, reputation, and branding.

Without in-person meetings, “human connection” is largely imaginary — even when real. Whether one’s long-distance romantic relationship is with a human or a chatbot is functionally a philosophical concern, if you can’t tell the difference. Increasingly, many people won’t care — or even want to know.

Making us human

Makerspaces are in-person experiences. You have to be in the room to cut the cardboard. But who is “you?“ One of the things that we keep hearing in the dialogue around artificial intelligence in education is that AI can never replace teachers. (Everyone in the room typically nods their heads in agreement.) But is that true? Will teachers never be replaced by AI? Let’s start with online classes. For the moment, let’s assume that all of the students want to be in the class and are not using bots or avatars. Every function of the teacher can now–today–be convincingly simulated by AI. The teacher‘s face can be generated in video and synchronized with a nearly flawless colloquial, idiomatic voice. AIs have been able to answer questions with a fair degree of accuracy, for years. That accuracy increases daily as models grow more sophisticated. The more specialized an area of discussion, the more reliable the AI responses.

Image from ChatGPT

It’s human nature to prefer that the entity one is dealing with be an actual human. Nobody likes dealing with AI customer service. There's something vaguely insulting about it. But what if we can’t tell the difference? Even if we know that we are dealing with an AI, if it’s sufficiently convincing, after a while, we forget and start interacting with it as if it were human. Have you said "please" or "thank you" to ChatGPT? People will anthropomorphize the heck out of almost anything. When the “thing“ is convincingly human, we will almost certainly emotionally process that relationship in a way that is equivalent to that of an actual human. People have already engaged in romantic relationships with AI’s. AIs have become celebrities with large fan bases. Human psychotherapists have migrated from the office to the screen. There have been experiments with AIs in that realm, albeit with significant legal indemnification.

Look, I love my teacher colleagues as much as anybody. I think that what they do in the classroom is magic. But there’s really nothing magic protecting teachers from being replaced by artificial intelligence. It won’t happen today, but it will happen very soon in online environments. In-person teaching, particularly in elementary and early childhood education, will take longer. It will require convincing robotic human analogues, which are likely decades away. Kindergarten and preschool teachers will be the last to go.

But we live in a pragmatic world. If something works, or if it makes money, it will be implemented. Unless we exert our collective will, the forces that drive education will be economic ones. Machines are cheaper than people and easier to manage.

• Artificial Intelligence and Relationships: 1 in 4 Young Adults Believe AI Partners Could Replace Real-life Romance - Institute For Family Studies
• Create AI Video from Pictures - VisionStory
• Meet the Spanish AI model earning up to €10,000 a month - euronews
• The (artificial intelligence) therapist can see you now - NPR
• Why AI Struggles With Education - Forbes

A changed mission

There have been quite a few articles written by educators, urging operational AI education as a way for students to remain relevant in the work world. By “operational“ I mean the training of people in how to use AI tools in their daily work.  While, some of these articles do open the dialogue a bit wider by giving a passing acknowledgment to larger societal effects, most make the same mistake. There is no way to ensure employability in a post-AI world. We can’t operate as if society still offers equal opportunity to those willing to put in the effort. Indeed, it never did. 

Image from Chat GPT

While learning to incorporate AI tools into one’s professional workflow will be mandatory for those remaining in the workforce, obtaining those tools will not be an assurance of landing a job. Remember, the fundamental reason for corporations to integrate AI tools is to reduce the number of workers.

While most AI tools are not currently able to replace workers 1:1, they can significantly increase the productivity of nearly every worker today. Increases in individual productivity enable workers to replace multiple colleagues. This will result in a reduction in available jobs that is so large that it will fundamentally change society.

The mission of education is no longer solely to create workers. Deciding what it will be is the challenge that we should be spending our time on.