An eye-catching profile picture, music tastes that match your own and no opinion on whether pineapple should go on pizza.

As far as dating app profiles go, it doesn't get much better than that. You swipe right or send a like, and put your phone away, hoping the interest might be mutual.

Lo and behold, barely an hour's passed and you've matched. The excitement is only equalled by the overwhelming fear of what comes next: starting a conversation.

The need to come up with flirty, funny chat can feel like a constant pressure. "I found myself overthinking what the opening line should be," says Neo Cheng, a health care worker and vlogger. "The more you think, the more you go into this downward spiral."

Having seen it handle politicians' speeches, school homework and even some journalism, the 33-year-old Canadian decided to see if ChatGPT could be his digital wingman.

Back in February, Sky News asked the chatbot for tips ahead of Valentine's Day. But the cheesy chat-up lines ("If you were a vegetable, you'd be a cutecumber.") and creepy photo comments ("Your smile is so warm and inviting!") got a thumbs down from our dating expert.

But things have changed since then, with AI developing faster than people swipe left on my Tinder profile and Chatgpt enjoying a substantial upgrade.

Read more:
How AI could change our reporting
The race to regulate artificial intelligence

The experiment

Neo got ChatGPT to write a 100-word Tinder bio, based on some information about him.

"An introvert with a kind heart," it wrote, who "loves trying new foods" and is "looking for someone to share laughs and adventures with". It carried on about how much Neo loved everything from summer to helping others. The tone was more soppy romance novel than snappy dating profile.

Asked for something more concise, it went with: "Health care pro, introverted Aquarius and amateur YouTuber. Summer-loving meat-eater with a weird sense of humour. Enjoys Radiohead, Coldplay, and Justin Bieber. Seeking a partner in crime for laughs and adventures."

Then on to the photos. The AI suggested using a "clear, well-lit headshot", an action shot, and a candid photo to go last. Neo went to his Instagram page to find photos that matched its ideas.

Once the profile was complete, each match received a response written by ChatGPT based on prompts Neo gave it about each person's profile.

Talking to matches

Anyone who's used chatbots will know they can often sound a bit formal and use flowery language.

Politeness can go a long way, though, and AI delivers that in spades.

"I can only imagine how tiring it can be to learn so much new information, but at least you're making progress!" was ChatGPT's response when one of Neo's matches revealed they'd started a new job.

The chatbot is also a stickler for correct grammar, punctuation, and capital letters - some way from how most people write online. It may give off a good impression, says Neo, but it sounds a little inauthentic.

Its chat-up lines remained cheesy: "Excuse me, but I think you dropped something. My jaw."

Told the match had responded with a "lol", ChatGPT proposed: "Glad to see I can still make someone laugh! Want to grab a drink and see if we can keep the laughter going?"

Read more:
How teachers are facing up to ChatGPT
Recruitment team unwittingly recommends AI for interview

Standing out from the crowd

ChatGPT may have been keen, but Neo's experiment was just that: an experiment.

The "what I'm looking for" section of his profile was set to "just new friends" and matches were told about the true nature of his replies before the conversation led to a potential meet-up.

Read more:
Most Tinder users 'already in a relationship'

But a recent survey by cybersecurity firm Kaspersky and dating app, Inner Circle, found more than half of single men really would consider using a chatbot to help talk to matches. And 51% of women said they would use it to sustain multiple conversations, just as university lecturer Owen is doing.

He lets ChatGPT do most of the work and adds "personal touches".

"I hadn't used dating apps for a while," explains the 44-year-old. "Finding time to have multiple conversations was harder than going on an in-person date.

"Coming up with new lines to open the door to possible dates can be draining," he says. "Using AI has helped eliminate the 'writer's block' that comes with app dating."

A new era of catfishing?

But to Jay Dodds, co-founder of dating app Bonkers, which has put user safety at its centre, even AI-generated chat-up lines are a step too far.

"I hate the idea" of using AI to find or talk to a match, Dodds says. "I'm a tech enthusiast and there is a place for AI, but when it comes to dating apps it's the worst possible idea.

"Even if you create a chat-up line, if it's not something you would naturally do, you're already putting a false sense of yourself across.

"We're about promoting safety, and no catfishing is part of that."

Kaspersky and Inner Circle's survey also raised concerns about a new era of AI-based catfishing, with 57% of respondents believing its use in an online dating setting to be dishonest.

Read more:
Martin Lewis on 'frightening' AI scam
Google creates AI comparable to doctors

'Do a sense check'

Dating coach Hayley Quinn is similarly cautious about AI's role in looking for love. She encourages people to do a "sense check" before putting any generated lines into practice, ensuring they are appropriate.

And for those worried about being on the receiving end, there are things to look out for.

Neo says from his experience, signs include responses in full sentences "where everything's perfect". "If people are responding in full sentences, full stops, capped up properly, it would make me a little paranoid," he adds.

Click to subscribe to the Sky News Daily wherever you get your podcasts

Indeed, his experiment left him less convinced of AI's role in a genuine attempt to date.

"If you're using AI, when you actually go on a date it's to your own detriment," he says. "There is no help, you're on your own: what are you going to say?"

As if dating apps weren't already loaded with red flags, from suspiciously low-res photos to avoiding questions, this brave - or terrifying - new era of AI may have thrown up another one: Who is even writing these messages?