The ascendance of AI

It’s the time of year when professors are grading papers and looking forward to a holiday break. But many students with take-home exams will face a perennial temptation in a new form. The advent of artificial intelligence technology with the ability to create coherent prose text is now available to the public, and the latest iteration to make headlines is called ChatGPT, developed by OpenAI, “an AI research and deployment company” with the mission “to ensure that artificial general intelligence benefits all of humanity.”

ChatGPT can respond to questions or reply to prompts, and the results are often surprising in their readability. Screenshots of responses (many of them humorous) have been making the rounds on social media, to nearly unanimous recognition that a sophisticated AI tool like this has revolutionary possibilities for communication in a digital age. These platforms will only grow in sophistication.

If we consider how teaching and student learning have developed in secondary and post-secondary education over the last two decades, for example, the reliance on digital technology has become ubiquitous. The pandemic only hastened the move to digital text.

The use of AI technology to enhance or even produce text has long been a topic of scholarly discussion and academic concern. There are potential commercial applications well beyond the educational space as well. It could well be that what are known as lorem ipsum generators for websites will be replaced by passable prose generated by AI tools, even if only for filler.

But as AI text development becomes more advanced and sophisticated, the ability to pass as human-generated content passes from the realms of technical academic jargon to the mundane and everyday. It will become increasingly difficult to determine whether the origin of a particular text is a human or a computer, or some combination of both, as AI can be used to generate a base text that humans later edit or adapt.

We are well beyond the infinite monkey theorem thought experiment. AI could plausibly produce something that at least passes for Shakespeare. Indeed, someone might someday try to pass off an AI-generated text as a newly discovered and long-lost manuscript by the Bard.

With the increase in data that we have seen in the digital era and the exponential growth of text and media that will come when AI is automatically generating more and more of what we see, hear, and read, the need for trust becomes all the more critical. As we have seen with recent revelations about social media platforms like Twitter, the curation of content is inevitable. The question is how transparent, inclusive, and responsible such curation will be.

And while we sort through all these amazing developments and develop the cultural and spiritual resources necessary to survive and flourish in an age of technological revolution, there are some steps we can take to mitigate the damage that is done even while we seek to benefit from such advances. Teachers, for instance, will have to grapple with the ease and accessibility of AI-generated content, which might mean relying on AI to detect the influence of AI. Likewise, plagiarism detection tools will have to include ways of identifying undue AI influence on papers.

Part of what will drive the use of AI-generated text, including spam messages and the equivalent in various digital spaces, is that there is little economic disincentive to prevent someone from flooding us with AI-created content. So, both in terms of quality and quantity, trust will increasingly require us to turn to alternative methods in digital and analog contexts.

But the numerical growth and qualitative improvements of deep fakes and AI-generated text will put a premium on verifiable and face-to-face interactions. Instead of take-home essays submitted through digital course management services, professors might need to revert to analog methods, like in-class blue book exams and verbal recitations and dialogue. It may be too early to proclaim the death of the take-home essay, but it is undoubtedly on AI-assisted life support. The psalmist warns us not to put our trust in princes, and we might as well extend that healthy skepticism toward all human institutions and authorities, especially those like digital media that can deceive with such ease.

As the political scientist James Patterson speculates, “Given that online stories can be taken down or edited without notice, print is arguably the best method to preserve free discussions.” And the Incarnation, which we celebrate at Christmas, shows us just how meaningful and significant real life is, especially when compared with virtual imitations and digital substitutes. We can be thankful that we do have a book, the Bible, that we can turn to for wisdom and comfort in times of trial—a real book that teaches us about true reality using real words. Reality matters.