The collaboration marks a major shift in digital retail, transforming conversational AI into a native commerce engine that touches consumer experiences, digital advertising, scientific research, and internal production.
The rollout integrates Maybelline’s Makeup Virtual Try-On, powered by L’Oréal’s ModiFace augmented reality (AR) technology, directly into standard ChatGPT threads.
Instead of switching to an external app, users will be able to upload a photo or use their camera directly within the chat window to view makeup shades mapped onto their own face while chatting about a desired aesthetic or skin tone.
L’Oréal’s beauty-tech services already see over 120 million global uses annually, but this integration scales exponentially across ChatGPT’s massive ecosystem, which OpenAI reports has reached 900 million weekly active users.
L’Oréal is also pioneering “AI-native advertising.” Its top brands, including Lancôme and Kérastase, are optimizing how they surface during organic product searches in the U.S. Simultaneously, brands like Garnier, CeraVe, and SkinCeuticals are participating in OpenAI’s exclusive ChatGPT global advertising pilot.
This strategy drops highly relevant brand interventions into conversational feeds precisely when users ask for beauty advice, capturing consumers at the peak moment of intent.
Beyond the consumer-facing tech, the partnership reaches into the scientific lab. L’Oréal is deploying GPT-Rosalind, OpenAI’s specialized life sciences model, to map the skin microbiome for its La Roche-Posay brand.
By leveraging AI to decode bacterial and cellular behaviors, L’Oréal intends to drastically accelerate the development of advanced skincare formulas.
Internally, OpenAI’s newest models are feeding into CreAItech, L’Oréal’s secure platform for creative teams. The system generates high-fidelity marketing visuals while protecting unique brand identities.
This builds on a heavy internal tech push: L’Oréal has already upskilled 73,000 employees in generative AI and rolled out internal tools like L’OréalGPT to streamline operations.
Asmita Dubey, L’Oréal’s Chief Digital and Marketing Officer, emphasized that AI now enhances every stage of the company’s value chain, supporting both consumers and employees.
By weaving its beauty ecosystem directly into the world’s most popular AI assistant, L’Oréal is demonstrating exactly how legacy brands can claim the future of AI-native retail.
The rollout integrates Maybelline’s Makeup Virtual Try-On, powered by L’Oréal’s ModiFace augmented reality (AR) technology, directly into standard ChatGPT threads.
According to analytics firm Sensor Tower’s State of AI Report 2026, ChatGPT’s market share fell to 46.4% by the end of May, down from more than 50% in January.
The decline comes as users increasingly turn to alternative AI assistants, with Gemini capturing 27.7% of the market and Claude reaching 10.3%. Other platforms, including Grok, Perplexity, DeepSeek and Meta AI, each account for less than 5%.
Despite the drop in market share, ChatGPT remains the world’s most widely used AI assistant, boasting more than 1.1 billion monthly users. Sensor Tower also noted that the chatbot became the fastest application ever to surpass one billion monthly users, cementing its position as the industry’s leading platform.
The report found that users are becoming more willing to switch between AI assistants based on performance, ecosystem integration and company decisions.
It noted that OpenAI’s agreement with the U.S. Department of Defense in February coincided with a measurable increase in app uninstalls, suggesting that “brand trust and values alignment matter to users, not just features.”
Meanwhile, Gemini has gained momentum through its integration across Google’s ecosystem, while Claude has built a strong reputation for productivity and is closing in on ChatGPT’s user-retention rate.
The broader AI market continues to expand rapidly. Sensor Tower estimates that users are on track to download nearly 2.3 billion AI apps and spend more than $4.2 billion during the first half of 2026, up from $1.83 billion during the same period last year. The report said the figures indicate the industry is gradually shifting its focus from rapid user growth toward monetization.
OpenAI has also expanded advertising within ChatGPT, with about 17% of daily users seeing ads by May. At the same time, ChatGPT is increasingly directing shopping traffic to retailers including Walmart, Target and Costco, highlighting the growing influence of AI assistants on consumer purchasing decisions.
Despite the drop in market share, ChatGPT remains the world’s most widely used AI assistant, boasting more than 1.1 billion monthly users.
As society demands faster and more complex computing power, the search for hyper-efficient alternatives has never been more urgent.
Superconductor materials capable of conducting electricity with zero resistance and zero heat loss have long been envisioned as the ultimate solution to this energy crisis.
However, their practical application has historically been restricted by the extreme, cost-prohibitive cold they require to operate and their frustrating vulnerability to magnetic fields.
A ground-breaking study published in Nature Communications by researchers at Chalmers University of Technology in Sweden offers a paradigm-shifting solution to these limitations.
Instead of trying to alter the complex, notoriously stubborn chemical composition of superconducting materials themselves, the research team shifted their focus entirely.
They decided to engineer the foundation beneath them. By making precise nanoscale modifications to the substrate material before depositing an ultrathin layer of a high-temperature cuprate superconductor, the team fundamentally reshaped the material’s electronic behavior.
Using a specialized high-temperature vacuum treatment, the scientists sculpted a microscopic pattern of ridges and valleys on the substrate surface, measuring less than one-millionth the thickness of a human hair. When the superconducting atoms settled onto this textured foundation, the nanoscale topography guided their structural arrangement.
This atomic alignment created a tailored electronic interface that effectively stabilized the material.
As a result, the ultrathin film maintained its robust superconducting state at significantly higher operational temperatures than previously achieved, while demonstrating a newfound resilience against disruptive magnetic fields.
This breakthrough addresses a major bottleneck in quantum device development and power grid innovation, where stray magnetic fields frequently collapse zero-resistance states.
By proving that subtle physical alterations at the substrate level can drastically enhance performance, the Chalmers University team has introduced an entirely new design principle for quantum materials.
Rather than hunting for elusive new chemical elements, engineers can now use surface architecture to unlock the full potential of next-generation, energy-efficient electronics, bringing us one step closer to a zero-loss digital future.
Microscopic Foundations Are Unlocking High-Temperature Superconductors.
The star collapses under its own colossal weight, crushing its matter into an infinitely dense, infinitely small point known as a singularity, the heart of a black hole.
However, singularities present a massive headache for physicists because they are places where the known laws of nature completely break down.
Now, a groundbreaking study by theoretical physicists Daniel Jampolski and Professor Luciano Rezzolla from Goethe University Frankfurt offers a radical alternative to this cosmic dead end.
Their new mathematical model, utilizing Albert Einstein’s equations of General Relativity, suggests that a collapsing star could trigger the birth of a miniature, expanding universe hidden inside itself.
This mind-bending theory relies on an exotic, hypothetical object known as a “gravastar” (gravitational vacuum star). From the outside, a gravastar looks and acts exactly like a black hole, possessing the same immense mass and crushing gravitational pull.
Internally, however, it contains no singularity. Instead, beneath a thin shell of ordinary matter, a gravastar’s core is filled entirely with dark energy. Because dark energy exerts a powerful outward, repulsive pressure, it perfectly counteracts the inward crush of gravity, halting the star’s total collapse.
While physicists have theorized about gravastars for 25 years, no one could explain how a dying star could actually transform into one.
Jampolski and Rezzolla finally cracked this mystery. Their model demonstrates that as a massive star collapses and reaches an unimaginable density, the extreme compression of matter triggers a phenomenon akin to a miniature “Big Bang.”
Deep inside the dying star, an infant universe is born. Driven by dark energy, this new interior cosmos rapidly expands, pushing outward against the collapsing stellar material. The result is a perfect, stable stalemate: the outer layers of the dying star form a protective shell around a freshly born, expanding universe.
Ultimately, this research provides a vital mathematical roadmap for exploring the unknown boundaries of space physics. It also paints a breathtakingly poetic picture of the cosmos: the death of a single star in the universe might not be an ending at all, but the precise moment a new universe begins.
New theoretical models suggest that stellar collapse can result in the formation of an interior, expanding universe rather than a traditional gravitational singularity.
In response to this shift, HC Solutions, an HR consulting and talent acquisition firm, is planning the Inzira AI Summer Camp, an initiative aimed at equipping secondary school students with practical artificial intelligence skills during the upcoming school holiday.
The organisers say the program is designed to introduce students to AI in a structured and guided environment, with a focus on aligning learning with real-world employment needs.
According to Jane Kalisa, Senior HR Consultant at HC Solutions, the initiative reflects the company’s broader vision beyond traditional recruitment services.
“HC Solutions is not just an HR firm. We also have a Department of Technology,” she said. “In this department, we will be providing AI training during the summer holiday. This camp targets students in Senior 4 to Senior 6.”
She explained that the goal is to keep students meaningfully engaged during the holiday while preparing them for future careers in a technology-driven economy.
“We are going to engage the students so they learn how to use AI and remain productively occupied during the holiday,” Kalisa said. “This will not only help the students, it will also help parents and the country at large. Once they complete the program, they will be able to apply this knowledge in their daily lives and future careers.”
Beyond training, HC Solutions also sees the initiative as a long-term bridge between education and employment, with plans to connect trained students to job opportunities in the future.
“As an HR firm, we are building a strong database,” Kalisa added. “When these students finish school, we will be able to connect them with clients who need professionals with AI expertise.”
From an industry perspective, experts warn that the impact of AI on the labor market is already visible, with companies increasingly seeking candidates who can work with automation tools and deliver faster, more efficient results.
Igor Kabirigi, Tech Lead at HC Solutions, said the demand is already reshaping recruitment expectations.
“It’s not only children who need this. The labor market is also being heavily affected by AI,” he said. “Every company we work with is now asking for new hires who can automate processes and produce reports in a short time using modern AI tools.”
He emphasized that early exposure is critical in ensuring students are not left behind as workplaces evolve.
“That’s why children matter. It is far better to train them while they are still in school so that when they enter the workforce, they already understand how the new world operates,” he explained.
The Inzira AI Summer Camp is scheduled to begin on July 13 at Kigali Parents School in Kimironko. The curriculum is built around current labor market needs and includes modules such as AI automation for business processes and prompt engineering, which focuses on how to effectively communicate with AI systems to generate accurate outputs.
Organizers say the program will also include interactive and practical activities designed to make learning engaging while maintaining critical thinking skills.
“Our lead coach has extensive experience and has implemented similar programs in Kenya and Nigeria,” Kabirigi said. “We have developed a safe approach to teaching AI that does not make students lazy and preserves their critical thinking skills.”
HC Solutions believes the initiative represents a model for future skills development programs linking education to employment readiness.
“We believe this will be the first of its kind run by a company with deep roots in the labour market,” Kalisa said. “It will bring a dynamic shift in how students see themselves now and in the future.”
The organisers also plan to expand the training beyond students, targeting working professionals who may need support adapting to AI-driven workplace changes.
“In everything we do today, incorporating technology makes work better,” Kabirigi said. “Finance uses it, health uses it, every sector benefits from technology to improve efficiency, save time, and reduce costs. After this program, we plan to scale it to the existing labour force.”
Rwanda recently strengthened its national focus on artificial intelligence through the approval of a new National Artificial Intelligence Agency by the Cabinet on June 8. The move is part of the country’s broader digital transformation strategy aimed at promoting innovation, strengthening AI governance, and ensuring the safe and responsible use of emerging technologies across sectors.
The new agency is expected to coordinate AI development, support capacity building, and enhance data governance frameworks while encouraging innovation in both the public and private sectors. It also reflects Rwanda’s ambition to position itself as a regional leader in responsible AI adoption, with a strong emphasis on ethical standards and inclusive technological growth.
The report shows that 1.6% of transaction attempts involving consumers in Rwanda were flagged as suspected digital fraud in 2025, down from 2.7% in 2024 and significantly below the global average of 3.8%.
While the decline suggests improved detection and stronger safeguards, analysts caution that fraud risk is shifting rather than disappearing.
“Despite recording lower digital fraud rates than global and regional averages, Rwanda is experiencing a structural change in fraud risk,” said Amritha Reddy, senior director of fraud product management at TransUnion Africa. She noted that criminals are increasingly targeting identity and trust at the earliest stages of digital interaction.
Money mule scams dominate reported losses
Among Rwandan consumers who reported losing money to digital fraud in the past year, nearly three in ten (29%) said the losses were linked to money mule scams — the highest single category of fraud impact in the country.
Other major fraud types reported include account takeover (24%), social engineering (22%), fraudulent credit card transactions (22%), and unemployment benefits-related fraud (22%). Third-party seller scams on e-commerce platforms accounted for 21%, while identity theft stood at 19%. Messaging-based scams such as smishing, phishing, and vishing each accounted for between 13% and 15% of reported cases.
Experts say money mule schemes are particularly concerning because victims are often unknowingly recruited to move stolen funds, making them a key link between consumer fraud and broader financial crime networks.
“Money mules often are the bridge between consumer fraud and broader financial crime,” Reddy said. “As these activities scale, isolated scams evolve into wider systemic risks.”
The report highlights a shift in when fraud is most likely to occur in the digital lifecycle. In 2025, Rwanda recorded its highest suspected fraud rate at the account creation stage, at 7.7%. This was followed by account login (1.6%) and financial transactions (0.5%).
This pattern suggests that fraudsters are increasingly targeting identity systems rather than attempting to breach transactions directly.
“Rather than attacking systems directly, fraudsters increasingly impersonate legitimate users, blending into digital ecosystems designed for speed and inclusion,” Reddy explained. “Once a compromised identity is onboarded, downstream fraud becomes significantly more complex and costly to prevent.”
Across sectors, retail recorded the highest suspected digital fraud rate involving Rwandan consumers in 2025 at 7.1%, followed by online communities such as forums and dating platforms (2.1%) and gaming or betting platforms (1.7%).
While retail remains the most exposed sector, the data also shows a sharp drop in attempted fraud volumes compared to 2024, including a 97% decline in retail-related attempts.
The findings also reveal that Rwandan consumers increasingly value digital safety when engaging online. Easy payment processes (73%), confidence in personal data security (70%), and simplified forms and applications (64%) were identified as the most important features influencing trust in digital services.
“Consumers are willing to accept friction when it clearly enhances protection,” Reddy said, noting that security is becoming a key factor in digital trust and brand choice.
Rwanda’s efforts to strengthen its digital ecosystem, including investments in digital public infrastructure under National Identification Agency (NIDA) and the country’s broader Vision 2050 development agenda, are expected to reinforce secure onboarding and electronic know-your-customer (eKYC) systems.
The introduction of a new digital identification number is seen as a key step toward improving identity verification and reducing vulnerabilities at the point of account creation.
Fraud evolving, not disappearing
Despite the decline in overall suspected fraud rates, experts caution that the threat is becoming more adaptive, subtle, and harder to detect. Fraudsters are increasingly relying on psychological manipulation and cross-channel scams that exploit trust rather than technical vulnerabilities.
“Fraud is becoming more automated, more targeted and more psychologically manipulative,” Reddy said. “The central challenge ahead is no longer simply preventing fraud but preserving trust in the digital economy.”
The report draws on a global survey of 12,730 consumers across 18 countries, including 308 respondents in Rwanda, conducted between November and December 2025, alongside TransUnion’s global fraud intelligence network.
While Rwanda continues to outperform global averages in controlling suspected digital fraud, the findings suggest that the next challenge will be addressing increasingly sophisticated identity-based scams at the earliest stages of digital engagement.
Analysts caution that fraud risk is shifting rather than disappearing.
The IPO values the company at about 1.77 trillion dollars, according to reports. SpaceX shares are expected to begin trading on the Nasdaq on Friday. In terms of fundraising size, the IPO would set a record for an initial public offering. Based on the valuation, SpaceX would rank among the top 10 most valuable U.S.-listed companies.
Founded by U.S. entrepreneur Elon Musk in 2002, SpaceX has businesses in space launch services, satellite internet and artificial intelligence. The IPO has drawn strong market attention, while the company’s high valuation has also prompted discussion among market watchers about its profit outlook and the sustainability of its future growth.
Reports said SpaceX’s listing arrangement breaks with some Wall Street conventions, including setting the IPO price in advance and reserving a relatively high proportion of shares for retail investors. Musk will retain strong voting control over the company after the IPO.
Given Musk’s large stake in SpaceX, the valuation could significantly increase the value of his holdings and make him the world’s first trillionaire.
Founded by U.S. entrepreneur Elon Musk in 2002, SpaceX has businesses in space launch services, satellite internet and artificial intelligence.
Researchers tested several AI models using a classic psychology exercise called the Stroop task. The test is designed to measure attention and self-control in humans.
In the task, people see color words like “red” or “blue” written in different ink colors and must name the ink color instead of reading the word. This becomes harder when the word and color do not match.
The AI systems performed well on short lists of words, but their accuracy dropped sharply as the lists became longer. Some models went from over 90 percent accuracy to below 20 percent in longer tests.
When matching and conflicting words were mixed together, performance dropped even more. Researchers say the systems often returned to simply reading the words instead of following the instruction.
The study suggests that unlike humans, who can usually control attention and ignore distractions, AI systems still struggle with this kind of mental control, especially in longer and more complex tasks.
The findings highlight that despite impressive abilities, AI still has limits when it comes to focus and resisting distraction.
A simple color-word test showed that today’s smartest AI models can unexpectedly lose focus and fall apart on longer tasks.Credit: Shutterstock
The new technology combines two types of propulsion that spacecraft normally use separately: chemical rockets, which provide strong bursts of power, and electric thrusters, which use less fuel and allow precise movements over long periods.
Traditionally, spacecraft need different fuel systems for each type of propulsion, making them heavier and more complex. MIT researchers have found a way to use a single fuel for both systems.
The fuel, known as ASCENT, was originally developed by the U.S. Air Force as a safer alternative to toxic rocket fuels. Researchers discovered that it can also power tiny electric engines called electrospray thrusters.
Electrospray thrusters work by using electricity to charge particles in a liquid fuel and then releasing them into space to create thrust. While they do not provide the powerful push of chemical rockets, they are highly efficient and ideal for long journeys and precise maneuvers.
MIT scientists tested ASCENT in these miniature thrusters and found that it performed just as well as fuels commonly used in electric propulsion systems.
The technology is now set for its first real-world test through a NASA-backed mission called Green Propulsion Dual Mode. A small CubeSat satellite equipped with both chemical and electric thrusters will use fuel from a single shared tank.
Researchers say the system could one day help small satellites travel much farther than they do today, including missions to Mars and the asteroid belt.
The technology could also improve Earth observation missions, allowing satellite operators to quickly or gradually reposition satellites depending on the task, such as monitoring storms or tracking climate-related events.
These four flight-unit electrospray thrusters were delivered by the MIT Space Propulsion Laboratory to NASA for the upcoming Green Propulsion Dual Mode (GPDM) mission. Credit: Amelia Bruno
The planned overhaul is expected to transform ChatGPT from a conversational chatbot into a broader AI platform that integrates coding tools, autonomous AI agents and third-party applications. The changes are reportedly set to begin rolling out in the coming weeks through updates to ChatGPT’s web and mobile platforms.
The move comes as the San Francisco-based company seeks new revenue streams ahead of a widely anticipated initial public offering (IPO). While ChatGPT has attracted nearly one billion users worldwide, many continue to use the service for free, prompting OpenAI to place greater emphasis on paid products and enterprise services.
According to the report, OpenAI is redesigning ChatGPT to highlight coding capabilities, image-generation tools and software developed by external partners such as Canva and Booking.com.
The company believes AI agents capable of performing tasks on behalf of users, including scheduling meetings, booking travel and completing workplace assignments, could become more valuable than traditional chatbots focused primarily on answering questions.
OpenAI has also reorganized several product teams under a unified leadership structure led by Thibault Sottiaux, formerly head of the company’s Codex coding product. The long-term goal, he told the Financial Times, is to develop a personal AI assistant that can support users across both professional and personal activities.
The company has reportedly seen strong growth in Codex, whose weekly active user base has risen to more than five million. Business customers now account for about 40% of OpenAI’s revenue, a figure the company expects to increase further this year.
OpenAI has not publicly announced the reported overhaul.
The planned overhaul is expected to transform ChatGPT from a conversational chatbot into a broader AI platform that integrates coding tools, autonomous AI agents and third-party applications.
