AI Has a Hotness Problem

Also: Microsoft sales beat estimates as customers prepare for AI rollout


AI-generated imagery leans towards producing conventionally attractive depictions, often sidelining diverse representations. This bias, especially observed in Microsoft's Bing Image Creator that utilizes OpenAI's DALL-E, may perpetuate unrealistic beauty standards. The global community grapples with AI regulation, with a summit at Bletchley Park poised to deliberate this, reflecting the escalating regulatory competition among major countries. Meanwhile, Japan's tourism sector adopts AI for multilingual communication, easing interactions with global visitors. Leading AI researchers urge more investment in AI safety, especially as rapid AI advancements outpace related safety regulations. Microsoft reports stellar quarterly results, bolstered by AI anticipation, while OCBC Bank implements a generative AI chatbot for its global workforce, streamlining numerous tasks. In other news, Tripadvisor and the AI chip startup, REBELLIONS, introduce notable AI innovations and investments.


  • 🔥 AI Has a Hotness Problem

  • ⚠️ The world wants to regulate AI, but does not quite know how

  • 🏯 Japan leverages on AI to ease communication hurdles

  • 🏛️ Governments, firms should spend more on AI safety, top researchers say

  • 📈 Microsoft sales beat estimates as customers prepare for AI rollout

  • 🏦 OCBC to deploy generative AI bot for all 30,000 staff globally

AI-generated imagery is increasingly skewed towards producing conventionally attractive or "hot" depictions of people, with a notable absence of average or diverse representations. Microsoft's Bing Image Creator, powered by OpenAI's DALL-E, has been at the forefront of this debate, though responsibility for addressing this "hotness problem" remains contested between the two entities. OpenAI acknowledges that their model often mirrors unrealistic beauty standards, potentially perpetuating harmful societal expectations. The issue's origins—whether in the training data or the algorithm's intrinsic properties—are still unclear. Some platforms, like Bing Image Creator, even produce attractive outputs when prompted to generate "ugly" faces. This rising trend in AI-generated beauty standards can shape perceptions of what's considered "normal," leading to heightened and unrealistic beauty expectations, mirroring issues found in edited magazine images and celebrity photos.

The global community is grappling with the challenge of regulating AI due to differing opinions on what, how, and by whom it should be policed. Set against the historical backdrop of Bletchley Park, the British government will host the "AI Safety Summit" on November 1st and 2nd, where a select group of world leaders and tech magnates will discuss this pressing issue. The urgency stems from AI's global reach, with large language models being easily accessible and potentially harmful if not properly governed. Regulatory competition, especially between the EU, the US, and China, is intensifying, with each vying for dominance in setting AI standards. Interestingly, many major tech companies, once opponents of regulation, are now its proponents, fearing the risks of unregulated AI. While the London summit aims to facilitate discussions, the hope is that it will lead to concrete steps in AI governance, though this is just the beginning of a long journey.

Japan's travel sector is turning to AI-based technologies like language processing tools to facilitate communication with international travelers amidst a rapid rebound in inbound tourism and a scarcity of English-speaking staff. Noteworthy developments include an AI-powered machine at Tokyo's Seibu Shinjuku Station, created by Toppan Printing Company, capable of interpreting speech in 12 languages, and a new service, ConcierGPT, by Kotozna Inc., which aids communication between hotel staff and guests. These initiatives not only bridge language barriers but also maintain a human touch in interactions, boosting Japan's appeal to international tourists.

Prominent AI researchers have recently advocated for a substantial allocation of AI research and development funds towards ensuring the ethical usage and safety of AI systems. In a paper released ahead of the international AI Safety Summit in London, they suggest that both governments and AI corporations should earmark at least one-third of their AI R&D budgets for this cause. The paper, penned by esteemed individuals including three Turing Award recipients and a Nobel laureate, underscores the need for a legal framework holding companies accountable for any preventable harm inflicted by their AI technologies. This initiative emerges amidst a legislative void, with the European Union's inaugural set of AI safety regulations still pending approval. The call for increased safety investments reflects a growing concern within the AI community, accentuated by the rapid evolution of AI technologies which starkly outpaces the development of corresponding safety measures.

Microsoft exceeded Wall Street expectations in its fiscal first-quarter results across all segments, with a reported revenue increase of 13% to $56.5 billion, driven by growth in cloud computing and PC businesses as customers anticipate utilizing its AI offerings. The revenue from Microsoft's Intelligent Cloud unit, housing the Azure cloud-computing platform, rose to $24.3 billion. The company's association with OpenAI and the anticipation of AI-driven products like ChatGPT stirred enthusiasm among corporate tech buyers. Notably, the earnings per share were $2.99, surpassing the estimated $2.65. Microsoft's AI-integrated products, such as the "Copilot" for Microsoft 365, are garnering attention, despite being in the pilot phase.

OCBC Bank has integrated a generative AI chatbot, OCBC ChatGPT, for its 30,000 employees across 19 countries. Developed in collaboration with Microsoft Azure, this bot aids employees in writing, research, and idea generation. The initiative followed a successful six-month trial, during which participants reported a 50% reduction in task completion time. The bot operates within the bank's private cloud, ensuring security. Besides being one of the global pioneers in large-scale gen AI deployment, OCBC uses the technology for tasks such as personalizing customer experiences, recommending stocks, and detecting fraudulent activities. Currently, the bank employs or tests four primary gen AI functions, including aiding coders, transcribing calls, extracting data from company documents, and summarizing reports. With AI currently making over four million daily decisions for OCBC, this number is projected to increase to 10 million by 2025 as gen AI assumes more roles.

🛠️ AI tools updates

Tripadvisor has unveiled AI-generated summaries to provide travelers with insights on key features of hotels they research. This feature, following the successful beta testing of an AI-powered travel itinerary generator, presents a concise overview of each property, enabling travelers to make informed decisions quickly. It's part of Tripadvisor's ongoing effort to meld AI with its platform, utilizing review and travel intent data to offer valuable guidance to users. This update, gradually rolling out in the US and Canada, showcases popular themes and details from past traveler reviews, enriching the user experience in discovering important hotel information.

💵 Venture Capital updates

South Korean AI chip startup, REBELLIONS, is in discussions to secure US$100 million in Series B financing to accelerate the development of its next-gen AI chip, potentially valuing the company at over US$500 million. The three-year-old company has engaged with potential investors from Japan, Hong Kong, Singapore, and Europe and aims to close the funding round by year's end. This move is part of a broader trend of increasing investment in companies focused on AI hardware, driven in part by the rise of AI technologies like OpenAI's ChatGPT. Rebellions faces competition from global players like Nvidia and fellow startups. It plans to produce its "Rebel" chip with Samsung's 4-nanometer technology and has previously secured backing from significant entities including KT, Korea Development Bank, Kakao, and Pavilion Capital.

🫡 Meme of the day

⭐️ Generative AI image of the day