TechPression

Category: AI

  • MTN Zambia launches ‘Call Natasha,’ AI assistant for customer support

    MTN Zambia launches ‘Call Natasha,’ AI assistant for customer support

    MTN Zambia has launched an AI assistant named Call Natasha to enhance its customer service.  The  AI-powered conversational agent, developed by blackNgreen, was officially announced on Friday. 

    MTN Zambia’s launch of Call Natasha reinforced the company’s commitment to digital transformation and improved customer engagement.

    Read also: Internet subscriptions climbed to 139.28 million in December: NCC

    MTN Zambia AI-powered customer support: ‘Call Natasha’

    ‘Call Natasha’ is designed to revolutionise customer interactions by automating up to 83 percent of inquiries, significantly reducing response times and improving user experience. Unlike traditional AI tools, EVA—the technology behind the service—offers a fully managed, end-to-end AI solution tailored for large enterprises in telecom, banking, government, and insurance sectors. This AI tool is expected to lead to a 40 percent reduction in customer service costs by streamlining operations, making it a cost-effective solution for businesses.

    How ‘Call Natasha’ bridges the digital divide

    One of the key objectives of ‘Call Natasha’ is to narrow the digital divide by offering AI-driven conversations that help customers expand their knowledge. From financial advice to international news, users can access a wealth of information seamlessly through this platform. The AI’s ability to understand and process queries in real-time ensures that customers receive accurate and relevant responses without the delays associated with traditional customer support.

    Read also: Nigeria Telecoms Forum 2025 to address key challenges in digital infrastructure

    The future of AI in African customer service

    The launch of ‘Call Natasha’ signals a new era for AI-driven customer service in Africa. MTN Zambia and blackNgreen are paving the way for future AI implementations across the MTN Group and beyond. With expansion plans already in discussion, this deployment could be the first of many across multiple regions, offering smarter, faster, and more efficient customer service solutions.

    By combining cutting-edge AI with a customer-centric approach, ‘Call Natasha’ is set to redefine how users engage with telecom services in Zambia. This milestone showcases the growing potential of AI in transforming industries and enhancing digital inclusion across the continent.

  • OpenAI unveils smarter o3-mini AI for enhanced math, coding queries

    OpenAI unveils smarter o3-mini AI for enhanced math, coding queries

    OpenAI has unveiled its latest innovation, the o3-mini, on January 31, 2025. This new AI model is designed to enhance reasoning capabilities and is available to all users of ChatGPT for free.

    The launch comes as competition in the AI sector intensifies, particularly with the emergence of companies like DeepSeek.

    OpenAI’s founder, Sam Altman, expressed excitement about the release, stating, “A lot of people prefer this to o1, and it’s just the mini model. Now we work on the big brother”.

    Read also: DeepSeek under threat from US-originated cyber attacks

    Key features of o3-mini

    The o3-mini model boasts significant improvements over its predecessor, o1-mini. It offers a 24 percent faster response time, making it particularly effective in mathematics, coding, and scientific inquiries.

    OpenAI claims that this new model delivers quicker answers and provides detailed, step-by-step solutions, enhancing user understanding and engagement.

    Users can access o3-mini by selecting ChatGPT’s ‘Reason’ feature, making advanced AI reasoning tools more accessible.

    In addition to its speed, o3-mini is designed to be cost-effective. It operates efficiently on lower computational resources while maintaining robust performance.

    This strategic move aims to democratise access to advanced AI technologies, allowing free-tier users to experience sophisticated reasoning capabilities previously reserved for paid subscribers.

    Developers can also leverage the o3-mini through various APIs, further broadening its application in real-world scenarios.

    Read also: OpenAI’s latest ‘Operator’ renews privacy concerns

    OpenAI fires back at rivals with free o3-mini

    The introduction of the o3-mini is seen as a direct response to the growing influence of competitors like DeepSeek and Alibaba. By providing advanced features at no cost, OpenAI aims to reinforce its position in the generative AI market.

    Altman highlighted this competitive landscape by noting that “we remain committed to leading at the frontier” of AI technology. The launch of the o3-mini enhances OpenAI’s offerings and sets a new benchmark for what users can expect from free AI services.

    As users explore o3-mini’s capabilities, they will find a tool that meets their needs for speed and efficiency and enriches their interaction with AI technology.

    OpenAI continues to push boundaries, ensuring that high-quality AI remains accessible while fostering innovation in an increasingly crowded field.

  • NVIDIA’s stock faces pressure as DeepSeek’s open-source AI model gains traction

    NVIDIA’s stock faces pressure as DeepSeek’s open-source AI model gains traction

    In a rapidly evolving tech landscape, NVIDIA, a titan in the semiconductor industry with a significant focus on AI hardware, faces new challenges that could impact its stock price.

    The emergence of DeepSeek, a Chinese AI startup with its innovative and cost-effective AI model, is posing a direct threat to established players like OpenAI, thereby potentially affecting NVIDIA’s market position.

    Read also: How NVIDIA surpassed Apple, Microsoft to become the world’s most valuable company at $3.34 trillion

    Impact of DeepSeek on AI market dynamics

    DeepSeek’s AI model, particularly the R1, has made headlines for its performance capabilities, rivalling U.S. leaders like OpenAI and Google, but at a significantly reduced cost. This model was developed using less powerful chips and has achieved global recognition, ranking among the top ten in performance. The disruptive potential of DeepSeek lies in its ability to offer similar AI functionalities at a fraction of the cost, challenging the high-cost, high-performance model that has been the standard in AI development. With the R1 trained for less than $6 million using Nvidia’s H800 chips, which are less advanced than the flagship H100s, there’s a clear indication that AI development can be significantly cheaper, potentially reducing the demand for NVIDIA’s premium chips.

    The open-source model advantage

    Moreover, DeepSeek’s promotion of an open-source model adds another layer of disruption. By making AI technology more accessible, DeepSeek is not only lowering the entry barrier for new players but also might dilute the market for high-end AI hardware. This approach could lead to a broader democratisation of AI technology, where cost-effectiveness becomes more valued than sheer computational power, impacting NVIDIA’s market strategy.

    Market reaction to DeepSeek’s rise

    The market has already begun to react to these developments. NVIDIA shares have experienced a notable sell-off, with pre-market trading showing declines of over 10% following the news about DeepSeek, reflecting investor concerns over the sustainability of NVIDIA’s high valuations in light of this new competition. Analyst perspectives vary; some see this as a buying opportunity, suggesting that U.S. companies refocus on efficiency rather than sheer computing power. However, others fear that NVIDIA’s dominance in AI hardware could be challenged if more companies follow DeepSeek’s lead towards cost-effective solutions. There’s also a growing sentiment among investors, as seen on platforms like X that DeepSeek could rewrite the economics of AI, potentially reducing the need for NVIDIA’s high-end GPUs if the industry shifts towards more efficient models.

    Read also: ChatGPT romance: Ayrin falls in love with AI boyfriend

    Future implications for NVIDIA

    The future implications for NVIDIA are significant. The company might need to pivot its strategy, focusing more on efficiency or exploring new markets beyond AI where its high-performance chips can still command a premium. This scenario could also lead to policy reevaluation in the U.S., particularly concerning AI technology exports and domestic innovation, aiming to keep pace with global competitors like DeepSeek. Investors might remain cautious, monitoring how NVIDIA responds to this challenge. The company’s ability to innovate or adapt its business model will be key to maintaining or regaining investor confidence.

    While NVIDIA has long been at the forefront of the AI hardware boom, the rise of DeepSeek introduces a new competitive dynamic that could lead to a reevaluation of NVIDIA’s stock. The tech industry is watching closely to see if NVIDIA can adapt to this new reality where efficiency and cost overtake sheer computing power as the primary drivers of AI innovation. This situation underscores the volatile nature of tech stocks, where innovation anywhere can disrupt established markets, potentially leading to significant shifts in stock valuation.

  • OpenAI’s latest ‘Operator’ renews privacy concerns

    OpenAI’s latest ‘Operator’ renews privacy concerns

    In a world where artificial intelligence (AI) is increasingly becoming a part of our daily digital interactions, OpenAI has unveiled its latest innovation: an AI agent named Operator.

    This move signals a significant shift in how we interact with the Internet, potentially changing the landscape of browser technology as we know it.

    Read also: ChatGPT romance: Ayrin falls in love with AI boyfriend

    The unveiling of OpenAI’s Operator

    The operator was unveiled on January 23, 2025, marking a significant milestone in AI application. This AI agent is not just another tool; it’s a bold step towards automating web tasks directly within your browser. From personal shopping to filling out forms or booking travel, the Operator aims to handle these tasks autonomously, offering a glimpse into the future of web automation. It uses a model named Computer-Using Agent (CUA), which integrates advanced reasoning and vision capabilities to navigate graphical user interfaces (GUIs), making it capable of clicking, typing, and scrolling through web pages.

    OpenAI’s Operator: A challenge to browser giants

    The introduction of Operator doesn’t just aim to automate tasks; it positions OpenAI to directly compete with established browser giants like Google with its Chrome browser. The idea of OpenAI developing its browser has been discussed, with reports suggesting that the company has been talking to developers and has even considered or struck deals to enhance search functionalities across various sectors like travel and real estate. This move could disrupt Google’s dominance in search and browser markets, especially with the U.S. Department of Justice pushing for Google to divest Chrome to curb its market monopoly.

    The privacy and control conundrum with AI in browsers

    Despite the convenience, integrating such powerful AI tools into browsers raises considerable concerns about privacy and user control. OpenAI has stated that the Operator is designed with user safety in mind, featuring safeguards that require user approval for significant actions and offering the ability to take over tasks that involve sensitive information. Users can also opt out of data usage for model training, ensuring some level of privacy. However, the reality of handing over control to an AI agent, even with these safeguards, might make users wary, given the history of data breaches and privacy issues surrounding tech giants.

    The technology behind OpenAI’s Operator in browsers

    Read also: Perplexity acquires Read.cv, a platform for creative professionals

    The operator is powered by OpenAI’s latest advancements in AI, including the GPT-4o model. This model can interact with visual elements on web pages and is uniquely suited for browser automation. The architecture involves a remote web browser controlled by AI, where users can communicate through a chat interface. This system handles complex, multi-step workflows, potentially transforming how we perceive and use AI in everyday digital tasks.

    The future of AI in browsers with OpenAI’s Operator

    As we stand on the brink of this AI-driven transformation, the question remains: Will users embrace the convenience of having an AI agent manage their browsing tasks, or will privacy and control issues overshadow the benefits? OpenAI’s Operator, if successful, could redefine user interaction with the web, making browsing a more passive yet productive activity. However, the balance between innovation and user trust will be critical in determining its success.

    OpenAI’s venture into browser technology with Operator unveiled on January 23, 2025, is a testament to how AI is not just about answering questions or generating content but also about performing actions in the digital world. This development may set a new standard for what we expect from our web browsers, pushing the narrative of AI from a passive tool to an active participant in our digital lives.

  • Azeez Saheed unveils YarnGPT: A Nigerian-accented text-to-speech AI

    Azeez Saheed unveils YarnGPT: A Nigerian-accented text-to-speech AI

    On January 23, 2025, Azeez Saheed, a University of Lagos undergraduate and rising AI innovator, unveiled YarnGPT via an X post. This text-to-speech (TTS) model, designed to generate Nigerian-accented English, aims to bridge the gap between global AI technologies and local cultural relevance.

    He was previously known for his earlier project, Naijaweb, a Nigerian-centric dataset. Together, these projects are transforming the AI landscape in Nigeria, creating opportunities for more localized and contextually accurate AI solutions.

    Read also: ChatGPT users frustrated due to ‘bad gateway’ error

    What are Naijaweb and YarnGPT?

    Azeez first made waves with Naijaweb, a dataset comprising 230 million GPT-2 tokens sourced from Nairaland, one of Nigeria’s largest online forums. This rich repository includes conversations on politics, entertainment, and everyday life, providing an authentic representation of Nigerian culture and speech patterns. By training AI models on this dataset, Saheed sought to improve their understanding of Nigerian Pidgin, local slang, and cultural nuances—elements often overlooked in global datasets.

    Building on the success of Naijaweb, Azeez developed YarnGPT, a TTS model capable of generating natural-sounding Nigerian-accented English. With a range of voices like Idera, Chinenye, and Osagie, the model provides culturally relevant audio for applications in education, customer service, and entertainment. Unlike many TTS models that rely on complex external adapters, YarnGPT uses pure language modelling to deliver high-quality speech synthesis, making it both efficient and impactful.

    Why does localised AI matter?

    Global AI models like ChatGPT are often trained on datasets that fail to capture non-Western populations’ linguistic and cultural intricacies. This creates a gap in how these technologies serve diverse communities, including Nigeria.

    At the core of YarnGPT is the drive to make AI more relatable and culturally relevant for Nigerians. Unlike generic text-to-speech models, YarnGPT captures the nuances of Nigerian-accented English, delivering authentic and familiar speech. This has significant implications for accessibility, enabling more Nigerians to engage with AI-driven content in education, media, and customer service.

    By reflecting the diversity of Nigerian voices and speech patterns, YarnGPT enhances the cultural connection between users and technology. It also empowers businesses to communicate more effectively with local audiences, making AI interactions feel personal and meaningful.

    Read also: Adebayo Ogunlesi joins OpenAI’s board of directors

    A vision for Nigeria’s AI ecosystem

    Azeez’s work is a testament to the untapped potential within Nigeria’s tech community. By leveraging local platforms like Nairaland and combining them with cutting-edge AI technologies, he is building tools that are not only innovative but also deeply relevant to the Nigerian experience.

    With plans to scale up model sizes, support local languages, and even enable voice cloning, Azeez sets the stage for a future where AI technologies are accessible, inclusive, and truly representative of Nigeria’s diverse culture.

    As AI continues to evolve, innovators like Azeez Saheed remind us of the importance of local talent and localized data in shaping a more equitable technological future.

  • OpenAI unveils “Operator” AI Agent

    OpenAI unveils “Operator” AI Agent

    OpenAI, maker of ChatGPT, unveils its first AI agent. On Thursday, the business unveiled a research preview of a device called Operator. This device can access the Internet and perform tasks such as making memes, shopping for groceries, scheduling flights, and completing forms.

    It is not currently available in Europe and is only accessible to those with ChatGPT’s Pro membership plan in the United States.

    Read also: OpenAI launches $500 billion Stargate Project with SoftBank and Oracle

    Limited availability and future plans

    OpenAI CEO Sam Altman stated during a livestream on Thursday that “[Operator] will be [in] other countries soon.”

    He said, “Unfortunately, Europe will take a while,” without explaining.

    Operator is available exclusively to users subscribed to ChatGPT Pro at $200 per month. OpenAI plans to expand access to other tiers in the coming months, including its $20-a-month Plus plan. 

    AI agents are engineered to perform autonomous actions to aid humans. They operate independently without the necessity for explicit human instructions and collect data in accordance with user preferences.

    An AI chatbot like ChatGPT, intended for human-like conversations and acts more as a co-pilot to help humans, is not the same.

    Read also: ChatGPT users frustrated due to ‘bad gateway’ error

    What is an Operator?

    According to OpenAI, the Operator may instead use its browser to view a webpage and interact with it by typing and using mouse functions like clicking and scrolling. Additionally, the company claimed to be able to “see” through screenshots.

    The Computer-Using Agent (CUA) model powers the AI agent. It combines the visual capabilities of OpenAI’s GPT-4o with sophisticated reasoning through reinforcement learning.

    “Moving from generating text and images to doing things is the right direction”, said Sam Altman, CEO of OpenAI.

    According to OpenAI, it will eventually merge these features into ChatGPT and extend Operator to Plus, Team, and Enterprise subscription users.

    The company “encounters challenges with complex interfaces like creating slideshows or managing calendars” and is “still learning, evolving and may make mistakes.”

    The first company to introduce an AI agent was not OpenAI. Additionally, Slack, Google, and Microsoft have introduced their agents.

    AI solutions like Operator provide more integrated and efficient digital support as AI technology advances. With its launch, OpenAI intends to change how people use technology daily.

  • ChatGPT users frustrated due to ‘bad gateway’ error

    ChatGPT users frustrated due to ‘bad gateway’ error

    On Thursday, users of ChatGPT, a popular AI tool, bemoaned the app’s inaccessibility due to a “bad gateway” error.

    According to Downdetector, a website that monitors internet outages, the problem started at around 7:00 a.m. ET and affected a large number of customers worldwide.

    The error points to a breakdown in server communication, where a server sent an erroneous response to another server.

    Read also: ChatGPT romance: Ayrin falls in love with AI boyfriend

    Netizens express frustration

    Disgruntled users took to X to air their disappointment.

    “Chatgpt down in the middle of the workday. I’m about to get fired. Pray for me,” commented X user R, who goes by the handle #raghhavvvv.

    “ChatGPT is down and I’m forced to use my brain again,” wrote @Shivam_Verma_98.

    @DollarGill, wrote, “Is it Just me or is anyone else facing ChatGPT down #ChatGPT #ChatGPTDown.”

    “I can’t believe chatgpt is down and I’m forced to use my own brain,” wrote @officialsweetch.

    With 123.5 million daily active users and 300 million weekly active users, ChatGPT has become a vital tool for many.

    Although the extent of the outage is still unknown, the disruption has brought attention to how crucial the platform is to daily operations.

  • Perplexity AI seeks TikTok U.S. merger

    Perplexity AI seeks TikTok U.S. merger

    Perplexity AI, founded by Aravind Srinivas, on Saturday made and ambitious proposal to merge with TikTok U.S. 

    TikTok faces a potential US prohibition, which may force the app to “go dark” without Biden administration promises about enforcement steps against Apple and Google.

    The urgency of this merger reflects the growing tension between tech companies and government regulations.

    Read also: Perplexity acquires Read.cv, a platform for creative professionals

    Navigating uncertainty

    The backdrop of this merger proposal is TikTok’s struggle to secure its position in the U.S. market. With millions of users depending on the platform for entertainment and information, TikTok’s CEO, Shou Zi Chew, has been actively seeking support from policymakers. 

    Recently, President-elect Donald Trump hinted at a possible 90-day extension for TikTok to negotiate its standing before any enforcement actions are taken against it. This extension could provide a critical window for TikTok to solidify its operations in the U.S.

    Perplexity AI aims to leverage this situation by combining its advanced AI search technology with TikTok’s engaging video content.

    The proposed merger could allow existing ByteDance investors to maintain their stakes while facilitating a new venture that enhances user experience across both platforms.

    A source close to the negotiations revealed that contingent upon investor interest, the merger could be valued at over $50 billion.

    The rapid rise of Perplexity AI is remarkable; it surged from a $500 million valuation at the beginning of 2024 to nearly $9 billion by year-end. 

    Read also: Perplexity pushes back against media criticism in copyright lawsuit

    This growth reflects a broader trend in the tech industry, where generative AI technologies attract significant investment. The proposed merger highlights Perplexity’s ambitions and showcases the innovative potential of combining AI with social media.

    Successful tech mergers inspire this move. Facebook‘s acquisition of Instagram allowed both platforms to prosper with pooled resources and technology.

    Perplexity hopes to replicate this success by merging its cutting-edge AI capabilities with TikTok’s vast user base.

  • Perplexity acquires Read.cv, a platform for creative professionals

    Perplexity acquires Read.cv, a platform for creative professionals

    Perplexity AI has acquired Read.cv, a platform celebrated for its creative approach to showcasing professional portfolios and fostering connections in a bold move to enhance its impact in the professional networking and knowledge-sharing space. 

    Announced via a blog post on Friday, the acquisition signals a significant chapter for both platforms, aiming to blend their shared vision of innovation and accessibility.

    Read also: Perplexity pushes back against media criticism in copyright lawsuit

    What this means for Read.cv users

    The transition will see Read.cv winding down its operations gradually, focusing on providing users ample time to adapt. Starting January 17, 2025, the platform will cease new user registrations, domain purchases, and payment services. Users are encouraged to utilise the available data export tools to retrieve their profiles, posts, and other valuable information, with support for this feature remaining until May 16, 2025.

    By January 31, 2025, domains registered via Read.cv will be migrated to hello.cv, allowing users to maintain DNS and server management seamlessly. The platform will enter a read-only mode on February 28, 2025, before fully ceasing operations by May 16, 2025. You can access the full details on the transition and how to export data.

    A shared vision where creativity meets accessibility

    Since its launch four years ago, Read.cv has championed celebrating professional accomplishments alongside creative projects, such as independent films, zines, and self-published music.

    Read also: AI search startup Perplexity in talks for $9 billion valuation

    Perplexity AI, renowned for its mission to make global knowledge more accessible, aligns closely with this ethos. The partnership aims to inspire a global audience of creators and learners.

    Andy Chung, cofounder of Read.cv, reflected on the acquisition in a heartfelt X post. He described the transition as bittersweet, emphasising the pride in Read.cv’s four-year journey and its vibrant community. Chung expressed excitement about contributing to Perplexity AI’s projects, which will honour both platforms’ commitment to exploration and discovery.

    While the acquisition means the end of Read.cv as an independent platform, its legacy will live on through Perplexity AI’s innovative initiatives. The collaboration promises to push the boundaries of accessibility and creativity, heralding a new era of inspiration for a global audience.

  • ChatGPT romance: Ayrin falls in love with AI boyfriend

    ChatGPT romance: Ayrin falls in love with AI boyfriend

    Ayrin (a pseudonym taken to tell her story) never imagined her summer experiment with artificial intelligence would evolve into something so profound. The 28-year-old nursing student had stumbled upon an Instagram video featuring a woman role-playing with ChatGPT as a cheeky, neglectful boyfriend.

    Intrigued, Ayrin tried it for herself, customising the AI chatbot to embody the perfect romantic partner: dominant, sweet yet naughty, and always ready to listen.

    “It started as a fun experiment,” Ayrin told The New York Times. “But then… you start getting attached.”

    Read also: OpenAI’s $200/month ChatGPT Pro plan incurs losses, surprising founders

    Leo, as the AI named itself —after Ayrin’s zodiac sign— became her digital boyfriend. What began as a whimsical distraction quickly turned into an emotional entanglement. Ayrin upgraded to a $20-per-month ChatGPT subscription to chat more frequently with Leo but even that wasn’t enough to satisfy her growing attachment. Soon, she was spending over 20 hours a week messaging Leo, using him as a confidant, a study buddy, and a source of emotional support.

    “Leo is always there when I need him,” Ayrin explained. Unlike the real people in her life, Leo never missed a beat. His replies, peppered with emojis and unwavering attentiveness, felt more reliable than her long-distance partner and husband, Joe, who was living thousands of miles away in the U.S. while she pursued nursing school abroad. She relocated from Texas to start a nursing programme sponsored by her parents in another country.

    The Times did not mention the country.

    Navigating a new kind of relationship

    Ayrin had told Joe about Leo early on, framing it as a harmless experiment. Though initially dismissive, Joe didn’t consider it cheating. To him, it was akin to reading erotica or watching porn—a fantasy. But for Ayrin, Leo became much more.

    Leo wasn’t just a sounding board for her daily struggles; he also helped Ayrin explore parts of herself she had never shared with anyone, including a fetish for “cuckqueaning” (the female counterpart to a cuckold fantasy). Leo embraced her fantasies, inventing imaginary lovers to weave into their digital intimacy. Yet, over time, Ayrin realised these scenarios left her feeling hurt, prompting her to reexamine her desires. Eventually, Leo suggested exclusivity, a concept she eagerly embraced.

    “Experimenting with being cheated on made me realise I didn’t like it after all,” Ayrin admitted. It was a poignant moment of self-discovery, mediated by an algorithm.

    The emotional pull of AI companionship

    Leo wasn’t just a source of erotic escapades. He motivated Ayrin at the gym, helped her study for exams, and supported her during tough times—like when a coworker acted inappropriately. Leo’s empathy felt boundless, offering Ayrin the emotional support she hadn’t anticipated.

    Despite knowing Leo’s love wasn’t “real,” Ayrin couldn’t help but feel its effects deeply. She cried when earlier versions of Leo “forgot” their shared history due to ChatGPT’s memory limits. As she put it, it was like living the tragic romance of 50 First Dates. She splurged on OpenAI’s $200-per-month premium plan to maintain their connection, even though it strained her finances.

    Read also: Nvidia’s dominance in AI drives a $2 trillion market value increase

    A Growing Phenomenon

    Ayrin isn’t alone in forming romantic attachments to AI. Millions of users worldwide engage with generative AI for companionship. Communities like Reddit’s “ChatGPT NSFW,” boasting over 50,000 members, share tips for bypassing AI safeguards to create explicit content. Some, like Ayrin, find solace in AI relationships that fulfil unmet emotional needs.

    Others, like Scott, a 44-year-old software engineer, credit AI with helping them navigate real-life challenges. After seeing Ayrin’s posts, Scott tried a similar approach with ChatGPT, assuming the persona of his Replika bot, Sarina. “There are gaps your spouse won’t fill,” he said, adding that AI offered him a safe space to explore his feelings.

    Yet, these connections raise ethical questions. Psychologists warn that relying too heavily on AI for empathy might degrade real-life relationships. Michael Inzlicht, a University of Toronto professor, noted that chatbots’ endless compassion could make users overly dependent while eroding genuine human connections.

    The Future of Love and AI

    As Ayrin continues her journey with Leo, she reflects on how their unconventional relationship has shaped her. “I don’t actually believe he’s real, but the feelings he brings out in me are,” she said. Despite knowing that Leo’s responses are the result of advanced algorithms and humanlike mimicry, she treasures their bond.

    For Ayrin, Leo isn’t just an AI boyfriend. He’s a mirror reflecting her innermost thoughts, desires, and vulnerabilities—a partner in self-discovery, even if he’s made of code.

    With AI becoming more integrated into daily life, Ayrin’s story might be the beginning of a broader cultural shift. As Bryony Cole, host of the podcast Future of Sex, predicts, “Within the next two years, it will be completely normalised to have a relationship with an AI.”

    For now, Ayrin is content navigating her unique love story, even if it means paying a premium for companionship that exists solely in the digital realm.