The first ever national AI Conference was held in San Francisco in September, 2023. Speakers from the most influential companies in the world shared their insights and progress on the topic of artificial intelligence (AI). I think that myself and fellow Missional Marketing colleague, developer Jason Collier, represented the only Christian organization in attendance. We listened to over 25 speakers covering all kinds of topics related to artificial intelligence. If AI is on your radar, and you love conferences, we’re sharing our key AI insights for churches, and how they may help you on your path. Catch Jason’s blog here.
Read on as I share seven things I learned at The AI Conference that may impact your church.
1. We’re Just Getting Started with AI
Think back to the 1980s and the world of the Atari 2600 and Pong. (If you weren’t born then, check out this video of what gaming was like then.) Back then, we marveled at the advanced graphics of our Atari 2600, while having no idea of what lay ahead.
That’s where we are now with artificial intelligence. Changes are coming fast. In a few years AI as it currently works will seem silly and outdated compared to what it will eventually be able to do.
It helps to realize where we are at this moment. It’s great to get excited about AI’s potential, but we should anticipate fast-paced advancements in the coming years. At the conference, one speaker shared how the number of Large Learning Models (LLM) has multiplied many times over the last six months. In the coming year he foresees that number jumping by 50x. As I mentioned, this is all just getting started.
2. We’re in the Wild West of AI Ethics
Right now, there are very few standards when it comes to the ethical use of AI. Multiple lawsuits are moving through courts across the world right now centering on the inclusion of copyrighted material in the use of generative AI – most specifically image use. LLMs only work when they’re trained on vast amounts of data. That training information needs to come from somewhere. But what happens when that data is copyrighted?
– The AI Act
The European Union is attempting to tame the wild west of AI with the AI Act. This act puts forth multiple safeguards including transparency requirements when generative systems like ChatGPT are used, and being transparent with the public about any copyrighted materials included during training.
– Ethical LLMs exist
Companies like Anthropic are leading the way in creating LLMs that ethically source data and are safe to use.The big boys like Meta, Google, IBM, and Anthropic are all building safe and harmless LLMs. However, it’s worth noting that some aren’t building their platforms with safety in mind. Consumers will need to learn how to discuss the ethics of any system they use in their daily work.
Here’s an important question to ask yourself: Is the LLM I’m using training itself on my data? If so, how? And if so, am I comfortable with that? Every AI system is built to learn – that’s what makes it an AI system. BUT, are they training themselves on whether or not you liked their response, or are they adding your input into their LLM’s overall dataset? Good things to keep in mind.
3. AI Will Eventually Disappear
One person I interviewed shared a fascinating viewpoint. He noted how artificial intelligence will soon be baked into every interaction online. In fact, it will be so central to how we use every electronic device that we’ll stop noticing it’s even there.
Microsoft is launching Copilot, an AI companion to Microsoft products. With Copilot, a person can search, paste, and generate new content from any saved Microsoft document. Google, Salesforce, and others are building similar experiences. Recently, my Google slide document asked me to describe the slide I wanted to create and it took a stab at creating it for me. Pretty soon, AI will define how we engage online all together.
4. AI should be Helpful, Honest, and Harmless
This statement came up again and again during the conference. It’s a statement that sets the benchmark for building large language models. First and foremost, models should be built to do no harm. Second, they should be honest and ultimately helpful to the user. The largest companies at the conference mentioned these three values many times.
These ideals are a great place to start. However, for Christians it raises more questions.
- Who gets to decide what’s “harmless”?
- What happens if someone eventually believes Christian thought is harmful?
– Christian Concerns of AI
Christians may fear that believing Jesus is the son of God and the way to heaven will eventually be deemed harmful. Luckily, our present AI models make room for distinct religious thought. Some models (like Bard), try to stay far away from anything controversial. While others (like ChatGPT) are willing to act as religious experts and speak on religious matters. This makes platforms like ChatGPT much more useful than Bard when investigating spiritual matters (at this time).
An AI speaking to spiritual matters may scare many Christians, fearing artificial intelligence will offer bad advice. However, when framing a question with a Christian persona, responses are surprisingly helpful. For example, you could ask, “As an evangelical pastor, explain how a person becomes a follower of Jesus.”
– Does AI Have a Problem with Honesty?
A really big challenge right now is helping LLMs to be honest. These systems are created to generate responses, whether or not those responses are true. A well-thought out but untrue response is called a hallucination. As researchers develop new AI models, they’re consistently figuring out how to help the models speak honestly on the data they know about and to not speak on topics they don’t have information on. It’s an ongoing challenge for these systems. As a result, consumers should always double-check their responses from generative models.
5. AI is More Technical Now, but Will Become More Accessible Over Time
As with the start of any new technology, some AI development tasks are more complicated to implement right now. Yes, anyone can easily use ChatGPT, Claude, or other AI snippet creating platforms. But it isn’t easy building a two-step process where one AI technology does one task while another does a separate task. That is, it isn’t easy right now. AI gets easier and less expensive to use every day. For example, AssemblyAI, an AI-powered transcription service, continues to lower their cost of transcription per hour. And new apps are being added daily to websites like There’s an AI for That.
This is good and bad news for the common consumer. It’s bad news because some of the tasks AI can do right now require a developer to help make them happen. It’s good news because developers from all over the world are figuring out ways to make those tasks easily accessible and inexpensive for the non-technical user.
One note of caution here. As you step into the world of generative AI, you’ll start to find this paid service that does task A and that paid service that does task B. Take your time in choosing an AI app. If you’re going to spend money on AI, prioritize it on the tasks that will help you save the most time. Many AI platforms offer a free “playground” that might address most of your needs.
6. AI Should Augment Humans, Not Replace Them
There’s a famous test invented in 1950 called the Turing Test. The test is a method of inquiry for determining whether or not a computer is capable of thinking like a human being. At its core, the test seeks to uncover if a robot can act so much like a human that it would be impossible to determine whether you were talking to a real person or a robot. One of the speakers at The AI Conference said this test is silly because it sets too low a bar. He noted AI can do many jobs better than humans. Therefore, he suggested we need a different paradigm altogether.
Instead, artificial intelligence should be used to augment human activity; to help humans do their jobs better, not replace them. So, whatever task you’re trying to accomplish, AI should be used to help you do that task better.
I’ve ready many articles voicing concern over using AI within the church setting for fear of taking away human agency. The fear here is that people will depend on AI more than the Holy Spirit, allowing AI to write spiritual content instead of a Holy Spirit-inspired human. Here’s the truth: A person will employ their personal agency to the same extent with or without AI. If you leaned deeply into the Holy Spirit as you created content before AI, you’ll do so to the same extent after AI.
If you think about it, AI can play the same role a Google Search does now. If I’m preaching on the book of Daniel, I can do a quick Google search on the topic and copy down points made from some of the most gifted preachers in the land. Regardless of where I gather my info, I have to choose whether or not I’m leaning into the Holy Spirit to determine what God wants me to say to my people, or not.
7. Generative AI is like a Calculator for Language
This was probably my favorite quote from the entire conference. Patrick McFadin, a VP at DataStax described generative platforms like ChatGPT and Claude, as “a calculator for language.” If that’s true, it will change everything – just like calculators did.
– Calculators Weren’t Always Allowed in the Classroom
When first invented, calculators didn’t receive a warm welcome in the educational environment. However, in 1975, the National Advisory Committee on Mathematical Education suggested that all students above 8th grade should have access to calculators for all class work and exams. By 1983, calculators were allowed on advanced placement tests for Calculus. By 1994 the SATs allowed calculators – often considered the “tipping point” where calculators in classrooms became standardized.
The use of generative AI in the classroom will take a similar path. Teachers can currently fight against the use of ChatGPT. However, soon enough teachers and students will find ways to use these platforms to help improve and check work. These next few years may be a little muddy as teachers and students figure out how students should use AI within the educational context. But, soon enough, teachers will require that all submitted work must go through AI before submission.
– How this Impacts the Church
Pastors, what if you had a secret machine that could improve everything you wrote? A machine that could help you write policies, emails, letters, social media, and even help you research and prepare your sermon? This machine is available – for free – and it’s called generative AI.
Some pastors may try to stay away from using AI for now. But here are two things you should know.
- AI will be built into everything you do soon, so you’ll be using it anyways.
- Choosing not to use AI is like choosing to take an upcoming calculus test without a calculator. Yes, it’s possible, but it’ll be much harder and take much longer to accomplish.
Instead, every pastor should view the growth of AI as a wonderful gift to help you do more, work better, and save time. Generative AI is the new language calculator for pastors that will help you do everything you do better.
David’s Final Take on The AI Conference
My favorite session was: “Lessons Learned while Building Anthropic’s LLM Product, Claude” by Benjamin Mann, Co-founder of Anthropic
My favorite speaker was: Bryan Catanzaro, VP of Research at Nvidia
A new AI technology now on my radar is: Tools that combine data from multiple places like Myscale
One tip for those going next year: Book your hotel early.
Will I attend again in future: Yes.
– AI and the Future of the Church
It was amazing to be at the first AI Conference and I left with a few big takeaways for the church. First, I was stunned by how fast this technology is developing. New ideas are coming to market almost every day. Second, many companies are considering the ethical impact of their product. I loved hearing all the companies expressing how they’re building ethics into everything they do. Third, there is no escaping AI. A person could currently say, “I don’t trust AI right now so I won’t use it.” However, soon this will be impossible. As AI is being built into every single thing we do online, it will soon impact how we use every electronic device and online experience.
– AI Insights for Churches: Takeaways from The AI Conference
As for takeaways for the church, there are multiple things to consider. How many people drive across the country (or across town) without using Google maps or another online map? The answer is probably very few. We’re in a moment where Generative AI systems like ChatGPT are the new Google maps. Pretty soon, every person will be using these AI systems for certain common tasks.
Church leader, in the same way you’ve probably grown accustomed to and dependent on using Google for your daily online searches, you should start using AI systems to help accomplish your daily tasks. Whether you need an image for an upcoming sermon series, edits on a church memo, or a suggested strategy to reach your city for Jesus, generative AI systems can become your best friend.
Missional Marketing has developed SermonSpark to become your pastor’s sermon-writing best friend. And many others are coming to market with other church task-helping AI solutions. It’s a great time to jump in and try using AI to help your church go further, faster.
We Are Ready to Help You
We hope these AI insights for churches have been helpful for you. If your church is looking to get started with AI but isn’t sure where to start, reach out and let’s chat. We are ready to answer any question your church has about AI and help your church take meaningful steps toward the future. You can also start using AI to help with your sermon prep. for free today by using SermonSpark.