Woman files for divorce after ChatGPT ‘exposes' husband's affair through coffee cup
Image: Feyza Daştan/Pexels
Artificial Intelligence (AI) is seamlessly integrated into our lives, whether it's managing work tasks, controlling smart homes, or offering advice on parenting, fitness, and even mental health.
Platforms like ChatGPT have become trusted companions for many, solving everything from quick research queries to complex emotional dilemmas.
However, what happens when we place too much trust in AI?
One woman in Greece found herself at the centre of a media storm when she used ChatGPT in an unconventional way, asking it to interpret the coffee grounds at the bottom of her husband's cup.
Known as tasseography, this ancient fortune-telling practice involves interpreting patterns in coffee grounds, tea leaves, or wine sediments to predict the future.
What followed was a dramatic chain of events that led to her filing for divorce, allegedly based on ChatGPT's "reading."
Video Player is loading.
Play Video
Play
Unmute
Current Time
0:00
/
Duration
-:-
Loaded :
0%
Stream Type LIVE
Seek to live, currently behind live
LIVE
Remaining Time
-
0:00
This is a modal window.
Beginning of dialog window. Escape will cancel and close the window.
Text Color White Black Red Green Blue Yellow Magenta Cyan
Transparency Opaque Semi-Transparent Background Color Black White Red Green Blue Yellow Magenta Cyan
Transparency Opaque Semi-Transparent Transparent Window Color Black White Red Green Blue Yellow Magenta Cyan
Transparency Transparent Semi-Transparent Opaque
Font Size 50% 75% 100% 125% 150% 175% 200% 300% 400% Text Edge Style None Raised Depressed Uniform Dropshadow Font Family Proportional Sans-Serif Monospace Sans-Serif Proportional Serif Monospace Serif Casual Script Small Caps
Reset
restore all settings to the default values Done
Close Modal Dialog
End of dialog window.
Advertisement
Next
Stay
Close ✕
Ad Loading
When fortune-telling meets AI
According to reports in the Greek City Times, the unnamed woman, a mother of two married for 12 years, uploaded a photograph of her husband's coffee cup to ChatGPT, asking it to analyse the patterns.
The AI's response was shocking: it suggested that her husband was considering starting an affair with a woman whose name began with the letter "E".
It went even further, claiming that the affair had already begun and that the "other woman" had plans to destroy their marriage.
For the husband, this was just another episode in his wife's long-standing fascination with the supernatural.
For individuals already inclined to believe in fortune-telling or the supernatural, AI might appear as a modern, more credible alternative.
Image: Lucas Pezeta/Pexels
He claimed on the Greek morning show To Proino that she had previously sought guidance from astrologers, only to later realise their predictions were baseless.
But this time, her belief in ChatGPT's "divine" analysis led to irreversible damage.
Despite his attempts to laugh off the situation, his wife took the accusations seriously.
Within days, she asked him to move out, informed their children of a pending divorce, and even involved a lawyer.
The husband, who denies any wrongdoing, is now contesting her claims, arguing that they are based on fictional interpretations by an AI and have no legal merit.
This bizarre story strikes at the heart of a growing concern: our increasing reliance on AI and its impact on how we perceive reality.
While tools like ChatGPT are designed to generate plausible responses based on the data they've been trained on, they don't possess true insight, intuition, or the ability to divine the future.
The AI chatbot, for instance, wasn't created to interpret coffee grounds or predict human behaviour.
It's a language model that generates responses based on probabilities, influenced by the prompt it receives.
Yet, incidents like this highlight how easily people can misinterpret AI's capabilities, especially when they lack a clear understanding of its limitations.
On Reddit, reactions to this story were a mix of humour and concern.
Some joked about AI taking over psychic jobs, while others expressed deeper fears about how tools like ChatGPT are blurring the line between reality and fiction.
One commenter aptly noted, 'We're going to see a wave of people whose ability to comprehend reality has been utterly annihilated by LLM tools.'
As AI continues to evolve, users must understand what these tools can and cannot do.
ChatGPT, for example, can provide thoughtful responses, summarise information, and even offer creative input.
But it is not omniscient, nor can it accurately predict human behaviour or events.
Moreover, this story underscores the psychological vulnerability that some people may have when interacting with AI.
For individuals already inclined to believe in fortune-telling or the supernatural, AI might appear as a modern, more credible alternative.
This misplaced trust can lead to significant emotional and relational consequences, as seen in this case.
As we continue to integrate AI into our daily lives, it's essential to remember that these tools are not infallible.
They are here to assist us, not replace critical thinking, human intuition, or professional expertise.
And perhaps most importantly, they are not fortune-tellers, no matter how convincing their responses may seem.
Try Our AI Features
Explore what Daily8 AI can do for you:
Comments
No comments yet...
Related Articles
Daily Maverick
3 hours ago
- Daily Maverick
AI job threats – graduates at risk, not call centre agents
As it turns out, South Africa's call centres are not a bastion for John Connor's resistance against AI. Instead, Judgement Day is coming for our graduates and young professionals. In the first three months of 2025 the graduate unemployment rate in South Africa increased by three percentage points over the final quarter of 2024. This means that, according to the latest statistics, 11.7% of university graduates are unemployed – just shy of the 11.8% peak in the same period in 2023 that Hannah MacGinty noted in a master's degree submission at Stellenbosch University. The entry-level job market is currently the front line of a war between humans and machines. Young people are pouring out of tertiary institutions, waving their degrees proudly, and AI is the Maxim gun. It's so bad that Geoffrey Hinton, the so-called Godfather of AI, told Steven Bartlett that mass job displacement is 'more probable than not,' advising young people to consider trades like plumbing on a recent episode of the Diary of a CEO podcast. Reports suggest that AI could wipe out half of all entry-level white-collar jobs in the next five years. Call centres remain on those high-risk lists, but are staging a fierce resistance. Augmentation, not annihilation Leaders in the country's R120-billion business process outsourcing (BPO) industry don't see AI as an existential threat, but as a tool that can make their companies more competitive and expand opportunities for young people if deployed with care. 'Our philosophy was never to say, oh, we see this as some massive disruptor that's going to turn the whole industry upside down,' says Mervyn Pretorius, group chief technology officer at CCI, the continent's biggest contact centre player. 'It should multiply your output. It should not be your output.' That's the 'human-in-the-loop' philosophy CCI runs on. His local CTO colleague, Cobus Pretorius, explains from the company's in-country base in Umhlanga that while generative models are exciting, they're simply not robust enough for the messy reality of serving global clients across dozens of platforms. 'You can't just plug an AI into an airline account that has 27 different systems and expect it to work. It'll break very quickly,' he says. 'Most of our calls are high-emotion, high-complexity,' Mervyn explains. 'That's where humans matter most.' Cobus adds that, far from cutting jobs, AI has been a 'net contributor' so far: 'It's creating demand for prompt engineers, reviewers, people building dashboards. We're training for roles that don't even exist yet.' What this means for you Youth looking to start studying or enter the job market need to review their skill set with an eye on the future. According to ChatGPT, these are examples of skills that are likely to be in demand: AI tools navigation: Understanding AI chatbots, speech analytics, and customer relationship management integrated AI tools. Data interpretation: Using AI-generated insights to anticipate customer needs or identify trends. Automation management: Knowing when to intervene when AI fails or escalates a query. A growing force Bruce von Maltitz, CEO of 1Stream, which builds customer experience (CX) technology for more than 100 local clients, echoes that view. 'I have not seen any decrease in volumes. None of our South African customers are cutting jobs because of AI,' he says. That's because, in his experience, when things get personal or complex (think health issues, travel bookings or big purchases) people still want to talk to people. 'The complex work is still going to the agent.' He points out that AI is not cheap to build. Large corporates are investing millions, and those projects 'are net creators of jobs because they need sophisticated teams of people.' His company is working to make those tools 'affordable in rands' so that South African firms can participate in the transformation. Von Maltitz describes CX as the entire journey of interaction with a business, from discovery to after-sales service. AI allows businesses to anticipate and smooth that journey. 'South African consumers expect more from technology. They're pushing the boundaries all the time,' he says. Arming young people The global BPO market is growing strongly, from more than $300-billion this year to a projected $800-billion by 2034. Instead of decline, the sector is shifting from cheap labour arbitrage to higher-value, AI-enabled services. South Africa is well placed. With 20 years of deliberate investment, the country has carved out a niche as a hub for complex, emotionally nuanced customer service. Industry leaders say that empathy, humour and cultural resonance give South African agents a premium edge over cheaper transactional markets. Government incentives and programmes like CCI's CareerBox NGO, which recruits and trains unemployed young people, reinforce the industry's social licence. So, if AI isn't the clear and present danger, what is? The consensus from local leaders is that the Keep Call Centres in America Act of 2025 is a far bigger threat to South African jobs than AI itself. The legislation, currently before US legislators, would: Create a public blacklist of companies that offshore more than 30% of call centre work; Penalise noncompliance with fines of up to $10,000 a day; Restrict federal grants and loans for five years for any company on the list; Require call centre agents to declare their location at the start of every call, with US customers entitled to demand a domestic transfer; and Force companies to disclose when AI, rather than a human, is handling their service. Von Maltitz says the impact could be profound, even if the law isn't fully enforceable. 'Americans and Europeans don't want to do these jobs. But if this passes, companies will have to retool and find an answer. Not because it's a better answer, just because they need an answer.' The reputational risk is just as serious. A US bank listed as outsourcing to Cape Town could face consumer backlash, no matter the quality of the service. Advanced strategy The call centre industry's response to AI may serve as a template for how other sectors can navigate technological disruption. By focusing on human-AI collaboration rather than human-AI competition, these companies are not just surviving the AI revolution – they're leading it. That makes the real question not whether AI will change the nature of work, because it already has. We should be asking whether other industries will follow the BPO sector's lead in treating that change as an opportunity for elevation rather than elimination. For young South Africans entering the job market the takeaway is that while jobs of the future will still require humans, they will need different skills. The key is ensuring that our education systems, our companies and our policies are aligned to make that future as inclusive as possible. DM
The South African
10 hours ago
- The South African
The best AI humanizer may hold the edge over AI detection systems
AI writing tools like ChatGPT, Jasper, and are now mainstream. They can produce full essays, marketing campaigns, or LinkedIn posts in minutes. But there's a catch: audiences, search engines, and institutions are increasingly aware of 'AI-like' writing patterns. Google's Helpful Content Update, for example, penalises low-quality, AI-stuffed articles. Academic institutions have to use tools like Turnitin to spot articles generated by a machine. Even the most casual of readers can tell when an article sounds flat, formulaic, or robotic. This shows people's mistrust because they read your work, but they wouldn't trust it. That's where Best Humanizer AI comes in. It isn't just a grammar fixer. It's built to transform AI-generated drafts into natural, reader-friendly writing that resonates with humans and passes AI detection tests. Whether you are a student trying to escape detection, a marketer chasing higher SEO rankings, or a freelancer scaling content production, an AI tool that humanizes text instructions has surely become a necessity. The process is straightforward, but each stage matters if you want polished, undetectable results. Before pasting text into Humanizer AI, clarify your intent. Writing for academics? Keep a professional, structured tone. Writing for a consumer audience? You'll want something friendlier and more conversational. Humanising AI isn't about random rewriting, it's about adapting tone and rhythm to context. Copy your raw draft (from ChatGPT, Jasper, etc.) and paste it directly into the Humanizer AI editor. Unlike many platforms, Humanizer AI requires no sign-up, subscription, or credit card. That makes it frictionless to use even on a tight deadline. This is where the magic happens. Humanizer AI runs your draft through a multi-layer rewriting process: It adjusts sentence length and rhythm for natural flow. It rephrases overly predictable phrases that AI detectors flag. It softens 'too perfect' grammar, adding the subtle quirks human writing naturally has. Humanizer AI doesn't just hand you rewritten text it also shows real-time detection scores across popular AI detectors (like GPTZero, Copyleaks, ZeroGPT, and Turnitin). This transparency is rare among humanizer tools and gives you confidence that your content won't be flagged. Step 5: Polish Your Work with Some Personal Although it has produced finesse, the tool will even shine better if you put in examples, anecdotes, or particularities related to your brand. This will strengthen the tool's authenticity, making space for your original voice. If it is a high-stake piece of content like your dissertation or a deliverable to a client, feel free to run this humanized output once more through the AI tool, once or twice. Each run makes it even tougher for the text to be flagged by AI detectors. Once satisfied, copy the content directly into your CMS, essay submission platform, or social media scheduler. You've just turned an AI draft into humanized content ready for real audiences. A humanizer AI tool isn't just a 'detection bypass.' Done right, it's a competitive advantage for writers, students, and businesses. Makes Content Undetectable AI detectors analyse patterns like uniform sentence structure, predictable phrasing, and unnaturally polished grammar. Humanizer AI introduces linguistic variety and more 'human-like imperfection,' dramatically lowering the chance of being flagged. Google prioritises content more and more, which demonstrates strong E-E-A-T (Experience, Expertise, Authoritativeness, Trustworthiness). More human content flows better and keeps readers engaged longer, and all of these would reduce bounce rates-positive indicators in SEO. In other words, humanized content ranks higher. AI text often lacks rhythm. It may use the same sentence length repeatedly, making it dull to read. Humanizer AI breaks up that monotony, creating dynamic, engaging text. This is critical in digital marketing, where attention spans are short. Editing AI drafts manually can take hours. Humanizer AI does the heavy lifting in seconds, giving you more time to focus on strategy, research, or adding creative flourishes. The versatility is unmatched: Students can polish essays and reports while reducing the risk of plagiarism flags. can polish essays and reports while reducing the risk of plagiarism flags. Marketers can scale blog and ad copy production without sounding robotic. can scale blog and ad copy production without sounding robotic. Freelancers can deliver client work faster and more reliably. can deliver client work faster and more reliably. Social media managers can craft posts that feel conversational and authentic. The best AI humanizer isn't just designed for students or academics trying to avoid Turnitin flags it serves a much wider audience. Humanizer AI helps bloggers and SEO professionals write long, keyword-heavy articles so that they can rank on Google without bringing on spam filters or penalties. It is also used by corporate teams for editing reports, proposals, and internal communications to be read smoothly, in professional tones, and by humans rather than machines. Transforms text created by AI into something that bears personality, nuance, and style, so it no longer reads like something outputted by a machine but as authentic human writing. For creative professionals like authors and copywriters, it's also very important for agencies managing many clients in the tool's ability to scale very quickly, while at the same time maintaining the quality and natural voice of every single project. AI Draft (Before) 'AI-generated text is becoming very common in marketing. Businesses may need to use tools to improve readability and avoid detection.' Humanized Output (After) 'From ad copy to entire blog posts, AI-generated text is everywhere. But here's the challenge: much of it still feels robotic. Humanizer AI steps in to bridge that gap, refining tone, reshaping structure, and helping businesses publish content that feels genuine while slipping past AI detectors.' The difference is clear. The second version is smoother, more engaging, and more trustworthy to readers and search engines alike. To get the most from Humanizer AI, treat it as your first pass, not the finish line. Start by layering in manual edits: add a quick anecdote, a data point, or an industry insight to anchor the piece in real experience and authority. Subsequent to being humanized comes search optimization-right targeting keywords which should be integrated naturally- not stuffing but clearly placed in headlines, intros, and some key sections. Keep brand voice front and centre: use the tool to refine the structure and rhythm of an article or by changing the tone so that it sounds unmistakably like itself. Just to be more sure, validate your output with several AI detectors aside from built-in scores, especially to high-stakes content. And when quality really matters-client work, academic submissions, executive comms-run iterative passes: humanize, review, lightly edit and humanize again to maximise authenticity and readability. AI writing is here to stay. But so are AI detectors and sceptical readers. The balance of speed and authenticity is no longer optional; it's essential. Humanizer AI delivers exactly that balance. This makes it possible for you to take advantage of AI efficiency in the production of content that passes for human-made text. Whether you want to evade detection by AI, boost your SEO rankings, or engage your audience with real writing, Humanizer AI is the best all-round solution out there. If you want to scale content without losing authenticity or risking detection, start experimenting with Humanizer AI today. The results will speak for themselves.
Mail & Guardian
13 hours ago
- Mail & Guardian
Robots and real voices at upcoming Singularity Summit
15+ New Speakers announced highlighting the power of women in AI and explores robotics shaping the real world As the Singularity South Africa Summit 2025 draws nearer, the programme continues to expand with the addition of new speakers and a strong focus on robotics, automation and women at the forefront of AI. Taking place at the Sandton Convention Centre on 22 and 23 October 2025, the summit will unpack how exponential technologies are shifting the boundaries of work, creativity and society, and what this means for Africa. Presented in partnership with headline sponsor Old Mutual and metaverse partner UBU, the event will feature local and global thinkers at the forefront of AI, robotics, leadership and ethics. In a programme designed to provoke, inspire and equip, the latest speaker additions reflect the depth and diversity of ideas that have the potential to future proof Africa. New perspectives and powerful voices Among the newly confirmed speakers is Norwegian AI Singularity expert Anita Schjøll Abildgaard, who will explore the realities of work and staffing in the AI era. John Sanei will follow with insights on The IRL Imperative: The Only Trend to Follow in the Age of AI , exploring how to stay human in an AI-shaped world. Jennifer Cheng Lo will share her expertise on investing in digital assets. Young South African innovator Bohlale Mphahlele, the inventor of a safety alerting earpiece for women, joins the line-up to speak about her invention. Aaron Frank will present on AI for the Physical World: Understanding Today's Autonomous Robotics and Mobility Revolution . Vivian Lan will provide a virtual introduction to exponential technologies to help frame the conversation. Shayne Mann will speak on leadership through invention, while Annie Blecher will address the awake mind, where inner calm meets peak performance. Award-winning South African poet and performer Lebo Mashile will host the summit and debut an original AI-assisted spoken word piece that fuses technology with live performance. Aaron Vaccaro, president of Singularity University, will share a forward-looking perspective on what is next from the global innovation ecosystem. Reece Meyer and Nadia Jaftha, in conversation with Marko Stavrou, will unpack the future of the digital creator economy. Jordan Crypt will address the future of digital finance, and Valter Adão will cover the future of work in today's rapidly changing economy. Meet Maximus: Robotics in motion This October, the summit gets a robotic upgrade. Attendees will get to meet Maximus, a humanoid G1 Unitree robot who will be present together with his robotic canine companions Murphy and Mavis. Their interactive demonstrations will demonstrate how robotics is transforming mobility and real-world problem solving. To see Maximus in action, click here: 'The world is not waiting for the future to arrive. It is being built through choices, tools and leadership that reflect a rapidly evolving reality,' said Mic Mann, Co-CEO of Singularity South Africa. 'This year's summit puts those ideas on stage and brings together people who are asking better questions and creating practical solutions.' 'AI is no longer confined to labs or code,' added Shayne Mann, Co-CEO. 'It is in the systems we build, the stories we tell and the tools we use. The summit is a platform for people who are doing the work to ensure technology serves a broader purpose, shaped by many voices.' A platform for transformation The Singularity South Africa Summit 2025 continues to position itself as a key meeting point for leaders across business, technology, policy and education. The inclusion of strong female voices in AI, alongside the rise of robotics and automation, reflects the broader commitment to relevance, inclusion and action. 'We keep saying women belong in technology, but belonging is not the point. Do not invite us in because we are women. Expect us because we lead, build, design and create. Expect us to rebuild broken systems, not accept them, because change is central to innovation and growth. That is why Singularity matters to Old Mutual. It is not just a sponsorship. It is a bold statement that we are investing in platforms that challenge the status quo, amplify diverse voices and drive real transformation. We are not here to observe the future. We are here to help shape it,' said Celiwe Ross, Director: Strategy, Sustainability, People and Public Affairs at Old Mutual. The 2025 summit will feature keynote sessions, immersive workshops, a curated expo and structured networking opportunities. Delegates can also engage with the event virtually through UBU's metaverse platform. For updates and registration, visit
