Choosing a name has always been one of the heavier decisions of early parenthood – not because it’s technically difficult, but because everyone has an opinion and every option carries some invisible weight. A family name that feels like an obligation. A favorite name you loved until your cousin used it first. Something that means something in another language you didn’t research. And now, sitting somewhere in that mix, there’s a new voice in the room: the chatbot.
Parents have been turning to AI tools for baby name suggestions for a couple of years now, but 2025 and 2026 brought the trend into full view. The debate broke wide open on the first day of 2026, when a Maryland couple made national news for exactly the kind of reason that splits the internet cleanly in half: they used ChatGPT to name their firstborn son.
Sarah and Stephen Winkler, as Futurism reported, explained their approach plainly: “We were looking up on ChatGPT boy names that go well with the last name Winkler. And then once we found Hudson Winkler, we’re like, ‘give us a good middle name.'” Hudson Oakley Winkler arrived just before 4:20 a.m. on New Year’s Day, becoming Carroll County’s first baby of 2026. The naming choice sparked immediate debate about using AI as a crutch for personal decisions. Author and NPR podcast host Linda Holmes wrote on Bluesky: “What amazes me is how quickly people are losing either the confidence or the willingness to complete basic human tasks.”
The Algorithm in the Nursery
The Winklers were not doing anything unusual by 2026 standards. They were just honest about it. New parents are increasingly turning to AI to help with all kinds of childcare questions, and a growing share report using tools like ChatGPT for parenting strategies. Baby names are one of the more obvious places that impulse goes. The appeal isn’t laziness – most parents who reach for a chatbot in those exhausted, still-deciding weeks are using it the way they’d use a friend who happened to be very well-read and had no stake in the outcome. You can tell it the last name is difficult to pair with soft endings. You can tell it you want something Greek but not overused. You can tell it to rule out anything that appeared in a Netflix thriller. It won’t pout.
Expectant parents may use baby name generators and AI tools to narrow things down, but they’ve yet to fully hand over the final decision to the machine. Parents who have tried roping ChatGPT into the process generally found that it helps reduce the field – but most felt no algorithm can replace the human instinct that goes into making a name feel right. AI as a shortlist generator versus AI as the author of a child’s identity are meaningfully different things, even if the practical gap between them is smaller than it sounds.
The Name AI Keeps Choosing
While parents have been using AI to brainstorm names across every category, the tools themselves have a tendency, and naming expert Laura Wattenberg spotted it. Wattenberg, an authority on naming who tracks patterns, data, and cultural movements in the field, anointed the name Elara the 2025 Name of the Year, explaining that the choice grew from a trend she began noticing across AI-generated writing, where the same name appeared again and again regardless of genre or audience. “The AIs love to create characters named Elara,” she told TODAY.com.
Elara has roots in Greek mythology, with sources variously citing meanings including “bright,” “shining light,” “hazelnut,” and “spear,” according to Nameberry. It entered the U.S. top 1,000 baby names and currently ranks #942. And yet it turns up in AI-generated math problems, in AI-written novels, in character-building exercises across every genre a language model touches. Wattenberg’s explanation: “The AIs are trained on human-generated content, and then they’re trained to please us. They try their hardest to give us exactly what we want.” Because no living person is widely associated with the name, it carries no baggage. “We don’t have an Elara we hate.”
That baggage-free quality is exactly what makes the name appealing in theory, and also what makes it slightly unnerving in practice. A name that has never belonged to anyone is a clean slate, yes, but it’s also a name that arrived without a story. The AI saturation became noticeable enough that one teacher reportedly added a line to assignment fine print: if the main character in a student’s work was named Elara, the student would lose 99 points. Elara had become a tell.
What AI Actually Tends to Suggest
The pattern Wattenberg identified with Elara is not a coincidence – it reflects something structural about how large language models produce names. AI tools tend to gravitate toward vowel-forward, rhythmically smooth names with mythology or nature adjacent roots and no obvious cultural controversy attached. When prompted for top baby names, AI delivers what you’d expect: trend-forward, vowel-heavy, softly modern names that feel optimized for Instagram birth announcements. They’re clean, stylish, optimized – trained, as one creative team put it, on “the last five years of Pinterest boards.”
Names like Nova, Isla, Orion, Luna, and Elara keep circling back because they check every box an algorithm can be trained to check: pleasant phonetics, cross-cultural accessibility, no famous villains attached, trending aesthetic cohesion. What they can’t account for is the grandmother who would have loved that name, the specific sound of it called across a school playground, the way it sits next to a sibling’s name or a family surname with three hard consonants.
AI tools also look backward. They are trained on existing data, which means they are very good at identifying what has been popular and synthesizing the patterns, but they cannot predict which name a kindergarten class in 2031 will have four of. The irony of using AI to find something unique is that every parent using the same tool with similar prompts is likely to land in overlapping territory. The algorithm optimizes for “feels distinct” rather than actually being distinct.
Where the Real Trends Are Going
Some of the names AI tools favor are already starting to appear in real birth data – whether because parents are taking the suggestions, or because AI and parents are both drawing from the same cultural well. BabyCenter‘s data, drawn from more than 350,000 babies born in 2025, found that while Noah and Olivia held the top spots, there were notable shakeups in the top 10 for girls. Eliana and Aurora entered the girls’ top 10 at positions 7 and 9 respectively, bumping out Ava and Luna. Both are exactly the kind of names AI tools recommend enthusiastically: vowel-forward, celestial or mythological in feel, smooth to say out loud.
The beautiful first names conversation has always tracked cultural mood, but what’s new is the feedback loop. AI learns from human writing, produces names that reflect our collective preferences back at us, and then those names become more visible, which makes them more culturally present, which makes them appear more in human writing, which gets absorbed back into training data. The archive never gets smaller, only larger. Elara is not the end of this cycle. It’s just the most legible evidence that the cycle exists.
BabyCenter also reported that the top 100 got a broader refresh, with new girl names like Juniper, Oaklynn, Sienna, and Eloise entering, while boys’ names like Malachi, Ali, and Arthur arrived, signaling what the organization called “a turn toward strong classics with spiritual and ancient roots.” That second category – the strong, ancient, rooted names – is where AI tends to be weakest. An algorithm can generate Elara. It is less reliably good at generating Malachi, because Malachi requires context, heritage, and the kind of meaning that runs deeper than aesthetic smoothness.
Read More: Influencer Breaks Down In Tears Over Criticism Of Her Baby’s Uncommon Name
The Name Is Still Yours to Give
There is something worth examining in the backlash to the Winklers – the reflexive alarm at the idea of a couple using a tool to help them choose a name. People have always outsourced parts of this decision. To baby name books. To the Social Security Administration’s annual list. To the Reddit thread they read at 2 a.m. To the grandmother who pushed hard for a family name. The difference with AI is that the outsourcing is visible and feels deliberate, whereas all the other influences operated as background noise we didn’t have to account for.
The name Hudson Oakley is, for the record, perfectly fine. It is not inspired. It is not strange. It sounds like someone who will play lacrosse and have good posture. Which may be exactly what the Winklers wanted, and they found it efficiently, and their son will grow into it the same way every child grows into whatever they’re handed on the way out of the hospital. The name isn’t the thing. The person becomes the thing.
What AI cannot do – and what no naming tool has ever been able to do – is tell you how it will feel to call that name across a quiet house at the end of a long day. No app gets to give you that. The feeling of it in your mouth, the way it sounds when you’re proud and when you’re exhausted and when you’re calling someone back to the dinner table for the third time – that is still entirely yours.
Disclaimer: This information is not intended to be a substitute for professional medical advice, diagnosis, or treatment and is for information only. Always seek the advice of your physician or another qualified health provider with any questions about your medical condition and/or current medication. Do not disregard professional medical advice or delay seeking advice or treatment because of something you have read here.
AI Disclaimer: This article was created with the assistance of AI tools and reviewed by a human editor.