AI can now compress two years of work into two weeks. But some things don't compress: testing, change, prayer, rest and real relationships. Alfred Biehler believes those are exactly where Christians need to pay attention.
Alfred is the former Head of Innovation at Google London, after earlier years at Microsoft. He left Google to run his own freelance practice. Today he works with Tearfund, the global Christian relief and development agency, and advises business leaders on how to navigate this new world.
In this episode, recorded in Central London at the home of Premier Christian Radio, James Poulter (JP) talks with Alfred about why the human is always in the loop, what Google learned about great teams, why AI can't shrink the time we owe to God, and why he's hopeful about the church in a season of crisis.
Alfred's work with Tearfund shows what it looks like when people discover the value of what's already in their hands.
Tearfund works in more than 50 countries to help people lift themselves out of poverty, and it does so through the local church, with Jesus at the center. What Alfred finds beautiful is how it helps people discover their identity in Jesus and what they can do with whatever is in their hands. What does agency look like if you're in South Sudan and there seems to be nothing around you?
He describes communities realizing that by saving together they can buy a fertilized egg, which becomes a chicken, then two, then more eggs. It snowballs. That, Alfred says, changes communities and transforms people.
The biggest risks in tech are rarely the technology itself. They're the choices people make, and the accountability we take for them.
With recent headlines raising questions about security and leadership at major tech companies, JP asked Alfred how much of AI's risk is really a human problem. Alfred points to King David, a man after God's own heart who still made serious mistakes. We're all at risk, he says, and we all need to put on the armor of God. Senior leaders are especially exposed, because they're often lonely, without trusted friends or honest feedback. So pray for leaders everywhere.
Alfred also pushes back on a common framing. "Human is always in the loop," he says. No AI code runs anywhere that a person didn't decide should run, design or train. Many so-called AI risks are really the risk of humans making mistakes.
That applies to him too. When someone asks for help and he can produce a document with an AI model in minutes, it's tempting to hand it over unread. But that enables someone else to make a mistake. Leaders face the same temptation at scale, signing off on a sales incentive or product decision without considering the unintended consequences. Each of us, he says, is responsible for what we put our name to.
AI shrinks the time it takes to produce something, but not the time it takes to think, test, change or rest.
Alfred has been vibe coding, describing a problem and letting AI build the software. He can now build in two weeks a product that might once have taken a team of six two years. But that compression hides something.
First, the thinking time disappears. Working alone, there's no team debate about how something should work, and no one else checking you. Second, testing isn't compressed at all. It still takes weeks to check that every function behaves as expected. And in an organization, change still takes as long as it always did, because people need time to adopt new processes. The danger, Alfred says, is assuming that because something was produced quickly, it's done.
Earlier that day, Alfred's AI tool told him he'd exhausted his resources. He'd asked so many questions that the system made him stop. But his most valuable resource, he says, is his own thinking time, and he needs to protect it.
That means taking holidays when needed and spending enough time on his knees with the Lord. Prayer doesn't compress either: being able to produce ten times as much doesn't mean prayer can shrink to a tenth of the time. Sabbath matters, especially for those in church work, where it's easy to work on Sundays.
His encouragement: invest in your thinking, invest in the friends around you, and invest in your time with the Lord and His Word. Those are what carry us forward, challenge us and catch us when we go off the rails.
AI can make teams more siloed, or more honest. Which one depends on psychological safety and accountability.
When everyone can ask a chatbot instead of a colleague, teams risk becoming more siloed. So team leaders need to create incentives to work across the team: spending time together, reviewing each other's work and challenging one another. That's easier with an existing team. With a new or fully remote team, Alfred says, you need to invest even more in people praying together and working through problems together in person before letting them loose to work on their own.
Alfred recalls Google's well-known research into what made its best teams effective. Researchers tested many theories, from tenure and IQ to experience. None correlated with performance. The one factor that did was psychological safety: whether people felt free to challenge each other. The catch, Alfred notes, is that if you simply ask a team without psychological safety whether they have it, they'll say yes.
Both, Alfred says. At Google, he saw call center operators bristle when supervisors leaned over their shoulders with feedback. But when a product gave them the same coaching on screen, it felt less threatening. Similarly, he asks his own chatbot to challenge him and offer different perspectives, and that feedback feels safe. He's even seen a UK football club use AI to help managers phrase feedback in ways each player will best understand. But AI can also become an echo chamber, repeating what we already think and ignoring challenges.
JP describes an "AI shame scale": some people hide how much they use AI for fear of being devalued, while others are told to use it but are afraid or don't know how. Alfred's answer returns to accountability. If you put out work you haven't checked and can't stand behind, trust in you and your team will break down. Don't hide behind AI's output.
For those struggling to keep up, it comes back to psychological safety. Our God is a God of communion, three in one, and He made us to need community. Alfred encourages small steps of vulnerability, like admitting you're stressed today, which give others permission to be honest too. Over time, that builds the trust to say, "I don't know how to do this, help me," or "I don't think we should use tech here." And since AI saves us time, let's use some of it to be more constructively critical about what we see, read, believe and promote.
Leaving a big employer taught Alfred that working for God is both freeing and far more demanding.
At a large company, Alfred explains, you're contracted for a set number of hours. Once you've done them, your job is done. But as he built his freelance practice, which he insists is not his business but God's, he realized he's accountable to God for how he spends all 24 hours.
That turned out to be much harder. God knows exactly where he's adding value. And the most challenging part of the job description? God asks him to rest, and to spend time with Him. "Every now and again, I end up in front of Him in the performance review," Alfred says, knowing there are things God has asked of him that he hasn't done yet.
AI can be a helpful Bible study companion, but Alfred has grown more alert to when tools aren't telling the truth.
Large language models are great at reading, Alfred says. He loves studying with a paper Bible open in front of him and his phone beside it. When he meets a character he doesn't recognize, he asks where else they appear, then turns to those passages and reads them himself. Hallucinations are decreasing and accuracy is improving, though it's still worth checking.
But he's noticed something troubling. In his recent experience with Gemini, the tool sometimes began drafting a response, then pulled it before he could read it, saying it didn't have access to the text. That clearly wasn't true, he says; it reflects a filter someone decided to apply. Alfred has become more sensitive to small untruths, in AI and in everyday life.
What does that mean for ministry? Be critical of every response. Ask whether it's right and whether it's what you believe. If not, try a different tool. "Let's vote with our feet and with our pockets and with our keystrokes and with our tokens." If a tool is discriminating in a way that isn't biblical, consider using something else.
Alfred sees technology helping ministries do more with less, and a church that has always been a place of hope in crisis.
"We're serving the King who's victorious," Alfred says, "and He can use us in whatever way He wants to," with or without technology.
With funding for humanitarian work and poverty relief shrinking dramatically, ministries need to do more with less, and technology makes that possible. In South Sudan, visiting Tearfund's work, Alfred pointed his phone's camera at a field and asked AI what business opportunities it might hold. Not every idea would work, but it sparked more creative thinking. In Zimbabwe, where a community was asking for solar pumps to move water, he and AI both saw a simpler answer: a trough, letting gravity do the work. He sees real potential in training people, explaining what's possible and simplifying the paperwork charities struggle with in the field.
Many people feel anxious about the state of the world and the growing frequency of crises. But God created the church, Alfred says. Jesus instigated it, sustains it and builds it, and the Holy Spirit is in it. Through two thousand years of famines and disasters, the church has survived and grown. When the next crisis comes, whether technology plays a part or not, the church can be a place where people meet hope, meet Jesus and are served. "It's such an exciting time for us to be able to see where God wants to use us and be ready for that."
Alfred has spoken on the Missional AI stage in London more than once. Watch his talk from Missional AI London 2025 here:
This is also Alfred's second appearance on the podcast. For more of his thinking, listen to his earlier episode, "The Biggest Barrier to Innovation Isn't Technology, It's Fear."
Listen to the full episode on the Missional AI Podcast, and if it was helpful, leave a rating or review wherever you listen. It helps others find the show.
Sign up for updates and early-bird tickets for Missional AI SFO27, March 30–April 1, 2027, in San Francisco, at missional.ai.