
UPP News: What's new in the AI field? (Part 13)
Date
September 26th, 2024
Reading Time
5 mins
Tags cloud
What's news
Subscribe to our newsletter and stay up-to-date with the latest AI news updates.
1. Meta Connect 2024: Meta AI upgrades, Ray-Ban Meta real-time video, and more.

September 25th, 2024
Source: TechCrunch
Meta Connect 2024, led by CEO Mark Zuckerberg, focused on advancing Meta's ambitions in AI and the metaverse. In a feature that will be familiar to anyone following OpenAI, Google, and Apple’s application of image-based searches, Meta showcased its AI’s capacity to respond to prompts based on images, as well as edit photos provided to the AI based on feedback given by the user. It can also share them on Instagram Stories.
In AI, Meta introduced a vocal interface for Meta AI across platforms like Messenger and WhatsApp, offering voice interactions with celebrity voices.
The Llama 3.2 AI model was released, capable of interpreting charts, images, and maps, though certain features are restricted in Europe due to regulatory constraints.
Earlier this month, the label noting whether content has been edited using Meta’s tools was shrunk down, Facebook will pitch you on AI content that it generates which users want to share or engage with. Meta also announced a test that would use its AI translation tools to take creator content, create a translated dub, and sync the creator’s lips accordingly. So far, the experiment is limited to creator videos made in the U.S. and Latin America, with translations between English and Spanish.
Additionally, Ray-Ban’s Meta smart glasses gained real-time AI video processing, live language translation, and music streaming integration, further enhancing their appeal as a future consumer device.
2. Ai2’s Molmo shows open source can meet and beat, closed multimodal models.

September 25th, 2024
Source: Allen AI
Ai2 released Molmo, a small, open-source multimodal AI model that rivals larger models like GPT-4 and Google’s Gemini in visual understanding. Molmo can analyze images, answer questions, and perform tasks like object identification with minimal data (600,000 annotated images). Unlike big tech models, Molmo is free, can run locally, and aims to empower developers without needing costly infrastructure. Ai2's CEO, Ali Farhadi, emphasized that Molmo shows small, efficient models can match the performance of much larger ones, challenging the dominance of big tech in AI development.
3. Reddit is bringing AI-powered, automatic translation to dozens of new countries.

September 25th, 2024
Source: Reddit
Reddit is expanding its machine learning-powered translation feature to over 35 new regions in Europe, Asia, and Latin America, aiming to attract more non-English users. This update allows both posts and comments to be automatically translated into a user's preferred language, facilitating seamless conversations between different languages within subreddits. Users can also view posts in their original language. The move follows Reddit's earlier introduction of translation for French posts and comes as part of its strategy to grow its user base and ad revenue. Reddit plans to roll out the feature soon in countries like Germany, Italy, and the Philippines.
4. Airtel deploys AI to curb spam, calls for regulatory oversight on WhatsApp in India.

September 25th, 2024
Source: Airtel
Bharti Airtel, India’s second-largest telecom provider, has introduced a free, AI-powered spam-detection system aimed at tackling the country’s rising issue of spam calls and messages. This system, which Airtel claims to be India’s first network-based AI-powered solution, will automatically alert customers in real-time about suspected spam calls and texts. The service is being rolled out for Airtel’s 387 million users at no additional cost, offering a much-needed defense against the overwhelming influx of unwanted communications. A survey by LocalCircles found that 95% of Indians receive spam calls daily, with many getting at least three per day.
Airtel’s new system is designed to process an enormous volume of communications—1.5 billion messages and 2.5 billion calls daily. The system employs dual-layer protection, filtering spam at both the network and IT system levels. Developed by Airtel’s data scientists, the AI algorithm analyzes caller patterns, frequency, and duration in real-time to flag potential spam. Additionally, it scans SMS content for malicious links, cross-referencing a database of blacklisted URLs to prevent phishing attempts.
Beyond just filtering, the AI solution can detect suspicious activities like frequent IMEI changes, which are often indicators of fraudulent behavior. India has seen a surge in spam calls that go beyond annoyance, with many impersonating government authorities, delivery services, and credit providers. In response to this, Airtel CEO Gopal Vittal also called for stricter regulations on messaging platforms like WhatsApp, which have seen a rise in spam messages targeting users.
5. OpenAI rolls out Advanced Voice Mode with more voices and a new look.

September 24th, 2024
Source: OpenAI
OpenAI announced the rollout of Advanced Voice Mode (AVM) to more ChatGPT users, specifically those in the Plus and Teams tiers, with Enterprise and Edu customers gaining access next week. AVM, designed to make interactions with ChatGPT more natural, now features a blue animated sphere instead of black dots, and users will be notified via a pop-up in the app.
AVM also introduces five new voices—Arbor, Maple, Sol, Spruce, and Vale—bringing the total to nine, with a nature-inspired naming theme. Sky, a voice showcased in the spring update, was removed after Scarlett Johansson alleged it resembled her own voice.
While ChatGPT’s video and screen-sharing capabilities have not yet been included in this rollout, OpenAI has enhanced AVM’s ability to understand accents and made conversations faster and smoother. Custom Instructions and Memory features have also been expanded to AVM, allowing for more personalized interactions. However, AVM is not yet available in several regions, including the EU, U.K., and parts of Europe.
6. Snapchat taps Google’s Gemini to power its chatbot’s generative AI features.

September 24th, 2024
Source: Snapchat
Snap has expanded its partnership with Google Cloud to enhance the generative AI capabilities of its My AI chatbot within Snapchat. Utilizing Google’s Gemini AI, My AI can now process various types of information, including text, audio, images, and videos. This integration allows users to perform tasks like translating images of street signs or identifying healthy menu options.
Initially launched in February 2023 using OpenAI’s ChatGPT, My AI has seen a significant increase in engagement—over 2.5 times within the U.S.—since incorporating Gemini’s capabilities. Snap CEO Evan Spiegel emphasized that this partnership reinforces Snap’s mission to empower users to express themselves and learn more about the world quickly and share experiences with friends.
This announcement continues a decade-long collaboration between Snap and Google Cloud. However, the enhancement of AI features raises concerns regarding the potential risks to users, particularly children, as the chatbot becomes more intelligent.
Tags cloud
Newsletter
DISCOVER MORE

ENTER YOUR EMAIL
YOU WANT TO...
Hanoi, Vietnam
Web3 Tower, No. 15, Alley 4, Duy Tan, Cau Giay, Hanoi, Vietnam

















































![[Recap] UPP Global Technology JSC Establishing Anniversary](/homepage/news-section/new-4.webp)

























































