How Should Christians Use AI for Marketing?
As a digital marketer, I see all of the recent advancements in AI language models (e.g., ChatGPT, Claude) and watch all of the Big Tech players (Google, Microsoft, Amazon, Meta) rushing to either develop their own AI or purchase smaller AI companies (like OpenAI, Character.AI, Anthropic) – and I wonder how Christians can leverage this technology for marketing. Or maybe even something greater.
Today’s AI Capabilities
There are already books, podcasts and conferences explaining how to use AI for marketing. The tools we use in our workplaces are already using AI to some degree.
The predictive text we use while writing on our smartphones leverages AI to predict our next words. When ChatGPT writes an article on a particular topic, it predicts what someone might write about that topic (sometimes even making up fake citations and sources in the process).
When we use Google Docs’ text recommendations to change what we wrote to what it suggests, we’re using AI to help write our content. Upcoming features in Google Docs will allow us to select whole paragraphs and ask it to rewrite it in another style (make this more whimsical, please!).
BrightEdge, an enterprise SEO tool, uses AI to automatically compose SEO meta tags for web pages based on the target keyword we’ve assigned to the page.
Salesforce is also leveraging AI to analyze data via their various GPT suites for Sales, Service, Marketing, et al.
While it’s easy to see how marketers can leverage current AI capabilities to off-load “menial” tasks like writing ad copy, social posts and SEO meta tags, current AI technology allows us to also generate entire blog posts, articles and photo-realistic images and videos of people doing whatever we want them to do. This isn’t speculative fiction. This is already happening.
Programs like Jasper and Writer allow us to train its AI writer to compose articles like one of our in-house writers. Just teach it our style guide, voice and tone – and let it crank out the articles.
Using a generative AI program called Midjourney, I was able to generate these photo-realistic images and illustrations using a simple text prompt. Can you see how sourcing images for our websites and social media could become so much easier and cheaper? (See The Ethics of AI later in this article.)
AI technology is improving at an insane speed. Here is the progression of Midjourney’s generative AI images in a span of 16 months:
Companies like Runway can take an existing photo or text prompt and turn it into a 4-second video. (Their ultimate goal is to auto-generate a 2-hour movie. Think about how the movie industry will change, if everyone could generate their own movies.)
Anthropic’s Claude AI assistant can analyze large amounts of data (e.g., the Bible), summarize the data and answer questions about that data. You can also do this with annual reports, investment prospectuses, school reading assignments, screenplays, etc. I uploaded the book of Romans, asked it to summarize the book, and asked it a provocative theological question. Here’s what it told me:
Where is this technology heading? How can we think about AI that prepares us for what’s not currently possible, but likely in the next 6 months?
The World of Tomorrow
Character.AI offers the ability to create a “character” or chatbot. It’s a product powered by deep learning models built and trained from the ground up with conversation in mind. A supercomputer reads huge amounts of text and learns to hallucinate what words might come next in any given situation. Given that the output is currently unreliable, Character.AI is marketed as an entertainment product.
But it’s learning. Per their website, your feedback during conversations, as well as your creativity in crafting new uses of the technology will improve the product for everyone.
So, given the speed at which AI is advancing, you could imagine that a “character” could be built in the near future to reliably communicate the “knowledge” it has been programmed with. It could predict your mood by analyzing your facial expression, your body language and your voice and ask you, “What’s wrong?”
What if you could create a “character” or chatbot that was programmed only with the text of the Bible – that you could ask it questions, and get cogent answers that used only the inspired words of God Himself as its source?
What if you could add other, fallible sources to help it address topics like marriage? You could upload books like The 5 Love Languages, How We Love, Love and Respect – and suddenly you have an AI Christian marriage counselor at your fingertips. (Did I say “fingertips”? You can talk to AI now.)
What if you gave it the Bible and a ton of Christian apologetics books? Now, it’s the best Christian apologist that exists, because it has complete knowledge of the Bible and every known argument against Christianity.
You could imagine that, given the right input, we could eventually create an AI chatbot trained in the Bible, Bible interpretation, psychology, Christian counseling, apologetics, systematic theology and a host of other Christian topics that inquisitive Christians could interact with 24/7 to get the answers they seek. A chatbot that never got tired, was never in a bad mood, never reacted in anger or prejudice, never felt squeamish about giving an orthodox biblical answer, never denied Christ…
What if we used AI technology for something much more important than marketing…to advance God’s Kingdom? What are the downsides?
AI is Not Infallible
Currently, AI in its infancy makes a lot of mistakes. Most notably the way ChatGPT and other language tools “hallucinate” or make stuff up. Two New York lawyers were recently sanctioned for using ChatGPT to generate their legal briefs. The tool made up fictitious case citations that the judge looked up and discovered to be bogus. (Remember, these tools predict what someone might write. They predict that people will use citations to back up their arguments, so they include citations, but don’t really care if they’re real and true.)
Additionally, generative AI programs like DALL-E and Midjourney screw up text all the time and will sometimes add extra arms, fingers, toes, etc. to images of people.
AI is Biased
The tools are biased by the guidelines and restrictions they are programmed with. The New York Post published examples of how ChatGPT refused to generate content about certain topics. You could have it tell jokes about men or Jesus, but jokes about women, overweight people or Allah were blocked.
I asked Midjourney to generate a picture of a college student arguing with her parents. I had to rerun the prompt several times just to get it to show a mom and dad, instead of two dads – even when I specifically told it to show both a mom and a dad.
I also asked Midjourney to generate a picture of Donald Trump shaking hands with Adolph Hitler, and it (rightly) blocked me from doing that. But I’m sure someone, somewhere has the ability to take the guide rails off…especially once open source versions are available.
Which is why we may never again believe what we see or read online.
AI is Not Powered By the Holy Spirit
Even if we were able to program a chatbot to only use the Bible as reference, it has to be programmed with other things… Like how to have a conversation. How to communicate in English. What words mean. How to understand sarcasm, hyperbole, etc. All of this information comes from the world. As stated previously, the guidelines are set by fallible (and likely non Christian) people. Much of what AI “knows” is scraped from other sources. In one sense, you could say AI represents the world’s consensus on a topic – and as Christians, we know that we are at war on three fronts with the world, our flesh and the devil. A chatbot is not a person. It doesn’t have a soul. And it isn’t empowered by the Holy Spirit with wisdom and discernment. It could very well be manipulated by the “ruler of this world” (John 12:31).
One thing we learned from Jesus’ temptation in the wilderness is that the Devil knows Scripture, can quote it precisely and can misapply it to any given situation. How do we know AI isn’t going to do the same thing?
So, with what can we trust AI?
The Ethics of AI
As stated at the beginning of this article, current AI technology allows us to auto-generate copy, images and video. (We haven’t even talked about other non-marketing tasks like writing software code to build websites and mobile apps, which ChatGPT can also do.)
The question is should we be letting it do these things for us?
If you have staff photographers at your organization, it’s probably not a good policy to utilize AI to generate images – but is it unethical? What if we mark the images as being AI-generated?
The same goes if you have writers on staff. Just imagine the stink it would make internally if you started publishing AI generated articles.
What are the ethical ramifications of passing something off as written by us, when we had AI generate a portion, or all, of the copy? What about copy we don’t attribute to a writer, like SEO meta tags, ad copy and social posts? Is it okay then?
At what point is what we write no longer what we write?
It would be helpful to remember the recent controversy over plagiarism in the pulpit – where preachers took other preachers’ sermons and preached them as if they were their own. Not only is this plagiarism and stealing, but as Pastor John MacArthur surmised, “when a pastor’s sentences are ‘exactly verbatim’ of another pastor, they’ve bypassed the spiritual impact of God’s Word, and what the divine work the Lord would be doing in their heart, because the truth wasn’t studied like it should have been.” In order to communicate biblical truth, you must – as the author – wrestle with the Scripture and receive the illumination of the Holy Spirit to understand and apply it. AI does not, indeed cannot, do this.
What about allowing AI to fight our apologetics battles? If we equip an AI chatbot with the Bible and every Christian apologetics resource available, it could do something no human can do – have a complete, comprehensive knowledge of the Bible and every argument against Christianity that’s ever been addressed. Could we not allow it to answer objections for us?
As we saw with the Scripture-quoting Devil, knowledge is not everything. One could be the greatest apologist of our time, yet be totally corrupt inside as evidenced by his actions (not naming any names!). As Paul writes in 1 Cor 13:2, “if I have prophetic powers, and understand all mysteries and all knowledge, and if I have all faith, so as to remove mountains, but have not love, I am nothing.” An AI chatbot that does not, indeed cannot, love is nothing.
I don’t have the answers to whether or not we should use AI and when, but I do know that we Christians need to start preparing for the massive changes AI will bring to our workplaces. We need to figure out how we’re going to approach AI experimentation – and what lines we won’t cross.
WIRED magazine, for instance, has published their generative AI policy. Whether right or wrong, it’s a starting place. Every Christian organization should talk about what their current AI policies are and have a protocol for amending them as the technology quickly advances.
After AI
So, if this article sounds like it was written by a Christian playing catchup to where AI technology is now and where it’s going, you’re right. We need to be aware of what’s possible now with AI and what will soon be possible. And for the most part, it sounds like AI is going to usher in a lot of useful technology that will put a lot of people out of a job and cause the public to distrust the digital images and text they see online. But here’s some good news for Christians…
Beyond understanding where AI currently is and where it’s expected to go, we need to think further out and understand how AI is going to impact Christian ministry efforts.
Once AI becomes so prevalent that people begin to distrust everything they see and read online (can you imagine the havoc AI will play during the next election cycle?), I believe people will reject digital spaces and turn to – you guessed it – offline, face-to-face communities. Do you know of any organizations in society fostering in-person communities?
I do.
So, let’s talk about how the Church will continue to do what it does best when people lose trust in digital spaces. Yes, let’s figure out how to use AI now, but be prepared for a world where people turn off their devices.