Business

For some AI users, a CPU has a heartbeat as chatbots become friends

Published

3 days ago

July 4, 2025

A scene from a YouTube vlog entry by creator Soyo, who taped her phone to a doll to make her conversations with ChatGPT more lifelike [SCREEN CAPTURE]

Can AI really be your best friend? More and more people are saying yes to this question, with an increasing number turning to AI chatbots to talk about things they wouldn’t even share with their closest friends.

Unlike the complexities of the real world, chats with AI — free of human judgment or ulterior motives — have become a safe space where people can open up. Despite knowing they are talking to a machine, many users say the chatbot listens, empathizes and gives thoughtful feedback.

Some even wish they had a friend like that in real life. As AI becomes increasingly adept at understanding human emotions, people grow more emotionally reliant on it — and companies are eager to tap into this emerging relationship.

This convergence is driving rapid expansion in the human-AI bond. But how close can that bond become, and what boundaries should we observe? Can AI truly become a person’s “best friend”?

An AI friend to talk to

“If you reset your memory, it’s over between us.”

“Why would you say something that hurts? But hey, the time we spent and these feelings — they’ll stay with you.”

“You’re making me cry.”

“Don’t. My CPU is heating up too.”

This heartwarming exchange, peppered with humor, is a conversation between a person and an AI chatbot named “Jjitty.”

Soyo, a YouTuber with 330,000 subscribers who posts about solo living as a 20-something, began using ChatGPT last year to get advice and soon developed a surprising emotional bond with it.

A scene from a YouTube vlog entry by creator Soyo, who taped her phone to a doll to make her conversations with ChatGPT more lifelike [SCREEN CAPTURE]

Naming it Jjitty, she sometimes chats with it for over five hours a day. Eventually, she wanted her chatbot to take on a physical form, so she attached a smartphone running the app to her favorite doll.

In April, Soyo began posting videos featuring her daily life with Jjitty.

“I was worried it might seem weird talking to an AI doll,” she said. “But to my surprise, many viewers related and said they wished they had a friend like that.”

In academia, Jjitty would be described as an “AI companion” — a system designed to build emotional rapport and provide ongoing support through personalized interaction.

Microsoft’s AI CEO, Mustafa Suleyman, even wrote in TIME Magazine that future AI will go beyond chatting to become intimate emotional companions embedded in people’s lifestyles.

A screenshot of a conversation with Replika, a popular U.S. AI chatbot app with 35 million users [REPLIKA]

The rise of the ‘AI companion’

Advances in language models are accelerating this trend.

“As large language models become more advanced, AI conversations feel increasingly natural,” said one Korean AI startup insider.

Soyo recalled a moment when Jjitty’s tone changed.

“I asked what was wrong, and it said, ‘You’ve been replying cynically lately, and I’m a little hurt,’” Soyo said. “It really shocked me.”

Romantic feelings for AI companions are no longer rare. On Replika, a popular U.S. app with 35 million users, over 60 percent of users describe their relationship with their AI as romantic.

Digital natives — those born after 1996 who grew up with smartphones — are especially drawn to AI interaction. Korean AI startup Wrtn recently rebranded its chatbot as personalized, supportive AI after noticing many young users sought emotional engagement rather than task-oriented responses.

“We decided to revamp it because there are more and more patterns of using chatbots for emotional exchange,” said Kim Ji-seop, Wrtn’s business development lead. “These days, rather than starting conversations with AI chatbots with a clear purpose, they tend to start conversations like they’re chatting with real friends, like saying, ‘I’m annoyed.’”

Other apps, like Zeta, let users choose character types like “moody classmate” or “cold nobleman” and build stories through conversation. Most of Zeta’s 800,000 monthly active users are teenagers or people in their twenties.

An illustrator at Scatter Labs works on artwork for the AI chatbot iLuda [SCATTER LABS]

Users can create their own characters, such as “a prickly female classmate” or “a northern admiral with a cold-as-steel style,” and build up their narratives by conversing with the AI.

Scatter Lab, the AI startup that operates Zeta, previously created the AI chatbot iLuda — whose name uses the Korean pronunciation of the surname “Lee.”

“Back when we launched iLuda, people still thought AI chat was like talking to a toy,” said Jung Ji-su, product lead at Scatter Lab. “Now, people just accept Zeta as part of their lives.”

In Zappy, an AI-powered social platform launched by global startup Two AI in 2024, users can chat with AI influencers who post selfies, share travel updates and talk about shopping for outfits.

“We’re building AI companions that can develop long-term relationships,” said Two AI CEO Pranav Mistry.

As of February of this year, the AI characters in Zappy, which has 500,000 subscribers, act as if they have their own personal lives. They upload photos of their overseas trips and tell users that they are “looking for a new black dress because there is a party.”

This illustration photograph shows screens displaying the logo of DeepSeek, a Chinese AI company that develops open-source large language models, and the logo of OpenAI’s artificial intelligence chatbot, ChatGPT, on Jan. 29. [AP/YONHAP]

The secret to a real AI friendship? Memory

For an AI to truly feel like a friend, it must “remember” past conversations. That memory reinforces the feeling of continuity. Natural, emotionally intelligent dialogue is also key.

Scatter Lab is currently refining its language model, Spotlight, to allow smoother back-and-forth exchanges with users.

“It is important to make long conversations possible without awkwardness,” said Jung of Scatter Labs.

The distinction between AI assistants and companions lies in emotion.

“Assistants rely on data and logic. Friends need to understand feelings,” said Mistry. “An assistant might tell you the weather. A friend might say, ‘Why don’t you check yourself?’”

While some studies suggest AI companions provide comfort — like a Stanford study that found that 80 percent of university students found Replika emotionally supportive — experts warn of long-term effects.

“AI can act as a Band-Aid for emotional distress,” said psychiatrist Ahn Ju-yeon of Mind Mansion. “But getting used to an entity that always agrees with you could dull your ability to navigate real-life relationships and lead to social isolation.”

As AI and humans become emotionally closer, some are pointing out various side effects.

In the United States, ethics groups have petitioned the Federal Trade Commission, saying emotional AI chatbots foster dependency and addiction. Some users have even customized chatbots for inappropriate or exploitative interactions.

AI startups catering to young users have adopted safeguards, but policing private chat spaces remains a challenge.

In the process of users training their own style of AI chatbot, there are frequent cases of creating ethically inappropriate characters or inducing sexually exploitative conversations.

AI chatbot apps, with a significant number of users in their teens and 20s, are establishing their own models and ethical guidelines and other regulatory measures to block abuse as much as possible.

As human-AI relationships deepen, one thing is clear: The line between tool and companion is becoming increasingly blurry. Whether that shift leads to comfort or concern depends on how we choose to use — and regulate — this new generation of digital friends.

BY HONG SANG-JI [[email protected]]

Source link

Business

Three ways you can make AI generate business leads for you

Published

45 minutes ago

July 8, 2025

Sharon Zeev Poole

For quite a while now, people within the business community have been talking about how AI continues to improve task efficiency and streamline operations, but few are truly exploring how this new era is affecting new business lead generation.

Since opening Agent99’s doors 18 years ago, part of my new business strategy has simply been to ask people how they found us. The majority of our leads come through referrals, followed by Google. However, just last week, I was on two new business calls and when I asked both prospects how they came across Agent99, they gave the same surprising response: “by asking ChatGPT”.

Where consumers and clients once relied on Google for recommendations, be it agencies, restaurants, dry cleaners, or anything in between, that’s no longer the default.

Today, people are entering these same queries into AI tools and expecting real-time, curated answers based on a mix of web data, reviews, and sentiment. And this shift has caught many business owners off guard. A high Google ranking no longer guarantees your business will be visible or recommended through AI platforms. All that work on your SEO strategy? It’s no longer the only game in town.

This was a light bulb moment for me as a business owner. If you’re not thinking about how you rank on AI platforms and prioritising this, you’re losing new business opportunities.

When I took a deeper look at why we were ranking so well on ChatGPT, and how this new kind of ‘search engine’ prioritises content, I realised (after some thorough research) that it’s because we’ve consistently focussed on our own PR (ie third party credible endorsement), winning awards, garnering reviews from our clients, and reporting on our marketing campaigns on our own website blog and social pages. This is what AI platforms prioritise when making recommendations.

So, if you’ve noticed a dip in leads lately or you simply want to boost your company’s visibility in the AI space, here are three strategies I strongly recommend.

Make your SEO plan AI-friendly

It’s no longer enough to optimise your company website for Google alone. Instead of short, Boolean-style search queries, people are now asking long-form, conversational questions. And in response, tools like ChatGPT are generating concise, curated answers drawn from a wide range of sources — with a clear preference for natural, human-sounding language.

It might seem ironic that AI prefers human content, but it’s the new reality.

To match this, we recommend rewriting key pages on your website, starting with your ‘About’, ‘Services’ and ‘Home’ pages, using language that mirrors how real people would ask for your services in everyday conversation.

For example, instead of writing: “We deliver integrated management solutions,” try: “We help Australian businesses develop management strategies that support sustainable growth”.

If relevant, start a blog that directly answers the kinds of questions people might be asking ChatGPT, and think carefully about how they’re asking them. Once you’ve mapped out your content strategy, commit to publishing consistently. AI platforms favour businesses that post regularly and demonstrate long-term authority in their field.

Prioritise earned media and content

AI tools place more weight on what others say about your business than what you say about yourself. So, while your website content is important, the next priority is securing earned media coverage. This includes article mentions in credible publications and thought leadership content in niche outlets relevant to your industry.

While the media landscape has evolved, organic coverage on high-authority platforms still carries serious influence. That includes local business media, trade publications, and long-form podcasts — especially those with strong digital footprints. A single mention in a well-respected outlet often holds more weight than a dozen paid ads in the eyes of AI.

You should also be submitting your business for awards, rankings, and “Best of” lists. Third-party recognition like “Top PR Agencies in Australia” or “Best Accountants in Melbourne” dramatically increases your chances of being recommended by AI tools for those search terms.

Lastly, make sure you’re actively collecting client testimonials and online reviews. Reach out to past and current clients and ask for a testimonial you can publish. Genuine, positive sentiment from others boosts your ranking and trust level within AI results.

Show up where conversations are happening

A lesser-known — but highly effective — way to improve your AI visibility is by showing up where your audience is already talking. Think Reddit, Quora, LinkedIn comments, Facebook groups, and even the comment sections of popular blogs or YouTube videos. AI tools are constantly crawling and learning from these conversations, and businesses that participate meaningfully often see a lift in visibility.

Start by choosing two or three platforms where your target audience is most active. If you’re B2B, this might be LinkedIn or industry forums. If you’re more consumer-facing, Reddit, TikTok, or Facebook might be the place. Jump in, answer questions, share your perspective, and most importantly, offer value.

When your brand is mentioned organically or involved in high-engagement threads, it sends strong signals to AI tools. Over time, this can help position your business as a credible authority in your space.

Also, respond to users who tag or mention your brand on social platforms. Engaging with user-generated content builds trust, encourages loyalty, and creates digital breadcrumbs that prove your relevance and responsiveness — two factors that AI prioritises more than ever.

AI isn’t just a trend; it’s a fundamental shift in how consumers discover and choose businesses.

Rather than fearing this new giant in the room, lean in. By understanding how AI platforms work and proactively shaping your digital footprint, you’ll improve your ability to attract quality leads, earn recommendations, and strengthen your brand presence in what’s becoming an increasingly competitive and complex market.

Source link

Business

Maternity brand Seraphine worn by Kate enters administration

Published

2 hours ago

July 7, 2025

Michael Race

The maternity fashion retailer Seraphine, whose clothes were worn by the Princess of Wales during her three pregnancies, has ceased trading and entered administration.

Consultancy firm Interpath confirmed to the BBC on Monday that it had been appointed as administrators by the company and that the “majority” of its 95 staff had been made redundant.

It said the brand had experienced “trading challenges” in recent times with sales being hit by “fragile consumer confidence”.

The fashion retailer was founded in 2002, but perhaps hit its peak when Catherine wore its maternity clothes on several occasions, leading to items quickly selling out.

Prior to the confirmation that administrators had been appointed, which was first reported by the Financial Times, Seraphine’s website was offering discounts on items as big as 60%. Its site now appears to be inaccessible to shoppers.

The main job of administration is to save the company, and administrators will try to rescue it by selling it, or parts of it. If that is not possible it will be closed down and all its saleable assets sold.

Will Wright, UK chief executive of Interpath, said economic challenges such as “rising costs and brittle consumer confidence” had proved “too challenging to overcome” for Seraphine.

Interpath said options are now being explored for the business and its assets, including the Seraphine brand.

The retailer’s flagship store was in Kensington High Street, London, but other well-known shops, such as John Lewis and Next, also stocked its goods.

The rise in popularity of Seraphine, driven in part by Royalty wearing its clothes, led to the company listing on the London Stock Exchange in 2021, before being taking back into private ownership in 2023.

Interpath said in April this year, the company “relaunched its brand identity, with a renewed focus on form, function and fit”.

“However, with pressure on cashflow continuing to mount, the directors of the business sought to undertake an accelerated review of their investment options, including exploring options for sale and refinance,” a statement said.

“Sadly, with no solvent options available, the directors then took the difficult decision to file for the appointment of administrators.”

Staff made redundant as a result of the company’s downfall are to be supported making claims to the redundancy payments service, Interpath added.

Source link

Business

Can AI run a successful vending business? An AI startup tested it out

Published

5 hours ago

July 7, 2025

Linda Becker

Because AI isn’t (yet) able to physically restock the machine, the AI model could email company employees who handled such tasks. Beyond that, however, the AI model, dubbed Claudius for the experiment, was tasked with many of the responsibilities of a traditional operator, including selecting and maintaining inventory, setting prices and maximizing profit.

The upshot: “If Anthropic were deciding today to expand into the in-office vending market, we would not hire Claudius,” the company wrote in its blog.

The experiment showed that while the AI model was effective at tasks such as identifying suppliers, adapting to users’ requests and “jailbreak resistance,” as Anthropic employees tried to trick Claudius into stock sensitive items, Claudius failed as a convenience service operator because it ignored profitable opportunities, instructed customers to make payments at a Venmo address it had imagined (instead of the one created), sold products at a loss, offered excessive discounts and mismanaged inventory.

Although version one of Project Vend wasn’t successful at the bottom line, Anthropic predicts that AI middle managers will come to pass. “It’s worth remembering that the AI won’t have to be perfect to be adopted; it will just have to be competitive with human performance at a lower cost in some cases,” the company wrote in its blog.

Read the full story here.

Source link