- AI SAMOSA
- Posts
- Audio-Jacking: A New Threat Using AI (and other exciting AI News)
Audio-Jacking: A New Threat Using AI (and other exciting AI News)
Stay ahead of the curve with our cutting-edge AI Newsletter! Subscribe now for a byte-sized dose of AI excellence delivered straight to your inbox!

Howdy, Awesome People!
Welcome back to AI Samosa, your favorite AI Newsletter!
If this is your first time, having you here is great! Be sure to sign up to stay updated with the latest happenings in AI.
AI Bytes 📰
AI-Powered Local Business Search on Google Maps
Google Maps has launched an experimental feature utilizing AI to aid users in discovering local businesses tailored to their needs. Leveraging large language models, the feature analyzes Google Maps' vast database of over 250 million places, including photos, ratings, and reviews. Through conversational search queries, users receive personalized recommendations for nearby businesses, events, restaurants, and activities, categorized with relevant photos and review highlights. Users can refine their searches with follow-up questions, and group dynamics are considered, enabling families to find child-friendly venues. Currently, in an early preview stage, Google is gathering feedback from a select group of Local Guides to refine the AI technology before a wider release. This innovation may significantly impact local search and customer discovery, potentially driving more qualified traffic to niche businesses and lesser-known attractions. Businesses may need to optimize their online presence to capitalize on conversational searches and leverage this transformative technology effectively. Read more…

Smartphone production slows due to AI Chip Demand
Huawei is facing production challenges as it prioritizes artificial intelligence (AI) chips over Mate 60 phone production due to soaring demand and manufacturing constraints. The company's reliance on a single facility for producing both Ascend AI and Kirin chips has resulted in lower yields, impacting Mate 60 output quality. Despite Huawei's silence on the matter, the situation highlights its struggle to rebound from U.S. sanctions and underscores the impact of restrictions on AI chip sales to China. The Chinese government's push to bolster computing power has increased demand for Huawei's Ascend series, particularly the Ascend 910B chip. Huawei's clandestine advancements in chip manufacturing, possibly aided by Chinese chipmaker SMIC, have facilitated the Mate 60's launch with domestically made 5G-capable chips. Despite production bottlenecks, the Mate 60 series propelled Huawei to reclaim its top spot in China's smartphone market, albeit causing delays for other products like the MDC 810 computing unit used in advanced driver assistance systems. Read more…

Microsoft Inks Deal with Semafor to Create AI-Bot Curated News
Microsoft and Semafor have joined forces to integrate Semafor's AI chatbot into news creation. Under the agreement, Microsoft sponsors "Signals," a breaking news feed providing diverse global perspectives. Journalists will write content, while AI aids in research. The collaboration emphasizes combating generative AI threats to media and leveraging AI to enhance reporting. Partnerships with journalist organizations ensure comprehensive coverage, while AI tools facilitate rapid multilingual search across news sources. The initiative reflects media's adaptation to technological shifts and commitment to quality reporting amidst evolving landscapes. Read more…

Audio-Jacking: A New Threat Using AI
IBM Security researchers have discovered a technique called "audio-jacking" that uses generative AI and deepfake audio technology to hijack and manipulate live conversations. In the experiment, the AI was instructed to process audio from two sources in a live communication and intercept the related audio upon hearing a specific keyword or phrase, manipulate it, and send it to the intended recipient. The AI successfully intercepted a speaker's audio when they were prompted by the other human speaker to give their bank account information, replaced the authentic voice with deepfake audio, and gave a different account number. The attack was undetected by the "victims" in the experiment. The researchers point out that this technique could function as an invisible form of censorship, with the potential to change the content of live news broadcasts or political speeches in real-time. Read more…

Role Of AI In Flood Forecasting
Google's efforts in using AI modeling for flood forecasting and its collaborations with various organizations to provide early warnings and support to communities at risk of flooding. These collaborations involve partnerships with GiveDirectly in Mozambique, UNOCHA in Nigeria, ICRC in Somalia, and monday.com in India. Together, they aim to provide flood forecasts and information to support anticipatory actions, prioritize humanitarian operations, and develop early flood warning systems. Examples include providing cash transfers to households in Mozambique, generating flood alerts in Nigeria, and supporting humanitarian work in Somalia. Notably, a collaboration with monday.com has led to the development of an early flood warning system for the Red Cross in Assam, India. Google is expanding its coverage on the Flood Hub to more areas and countries, continually improving functionality, and adding more flood information to aid vulnerable communities. Read more…
This week’s How to Guide 🤯
Sure, here's a simplified 5-step process for using Midjourney' s Style Reference feature:
1. Create a Base Prompt: Write down the initial details of your image prompt, including the subject's description, clothing, atmosphere, and any other relevant details.
2. Generate the Base Image: Use Midjourney to generate the image based on the prompt you created.
3. Copy Image URL: Once the image is generated, copy the URL of the generated image.
4. Start a New Prompt: Begin a new prompt in Midjourney where you'll iterate on the details or create variations of the base image.
5. Use Style Reference Feature: Add "--sref" to the end of your prompt, followed by a space, and then paste the copied image URL. This tells Midjourney to use the style of the referenced image for the new prompt.
By following these steps, you can create multiple variations of the base image while maintaining consistency or style cues from the original image.
Stay tuned as we explore how this remarkable feature can be applied across multiple use cases.
TechCanvas 🎨

Narrative Nook📔
AI-generated story
Brushstrokes of the Soul
In the bustling streets of Tokyo, Hiroshi, a struggling artist, painted the city's heartbeat. Critics declared his work lacked depth, but he saw beauty in the mundane. One evening, a gallery visitor lingered, tears in her eyes. His art had mirrored her own story. Hiroshi's canvases ignited a movement, proving that every day can be a masterpiece of the soul when seen through the lens of passion.
Mind Teasers: Trivia Time! 🕵️
Who is the co-founder of Microsoft Corporation and is widely regarded as one of the pioneers of the personal computer revolution?
A) Bill Gates
B) Steve Jobs
What is the primary purpose of the programming language Python?
A) Web development
B) Data science and artificial intelligence
Which science fiction author is best known for his influential works, including "Neuromancer," and is often credited with coining the term "cyberspace"?
A) Philip K. Dick
B) William Gibson
Which AI technique, inspired by the human brain's neural structure, is used for tasks like image recognition and natural language processing?
A) Machine Learning
B) Deep Learning
Which AI milestone, achieved by DeepMind's AlphaGo in 2016, demonstrated an AI's ability to defeat the world champion in the board game Go?
A) Turing Test
B) AlphaGo Challenge
Check out the answers at the end of this newsletter 🔚 ⬇️
Tool Spotlight 🛠️
Effortless Video Editing with Unscreen

If you're a creator, you know that editing can often take twice as much effort as shooting a clip. While editing remains at the heart of bringing your visions to life, it also demands significant production value. However, if you're just starting and working with a tight budget but overflowing with creativity, Unscreen is the tool for YOU.
Unscreen effortlessly removes the background from your videos, eliminating the need for tedious chroma keying or pixel selection. Simply record your footage anywhere, and let Unscreen handle the rest. With its automatic analysis, Unscreen consistently delivers high-quality results without any manual input from you. It's amazingly accurate and incredibly user-friendly, making it the ultimate solution for seamless video editing.
Moreover, Unscreen Pro offers plugins for Adobe Premiere Pro & After Effects, along with support for various other editing software. So, if you're ready to take your editing game to the next level without breaking the bank, Unscreen has got you covered.
Editor's Note ✍️
This newsletter comes straight from the desk of John V Jayakumar, CEO of Superposition Technologies.
My life's journey is a testament to the transformative power of education and technology—my dad broke free from poverty through education, and I achieved millionaire status thanks to technology. My takeaway for you? Wholeheartedly embrace technology; it's a game-changer for you and everyone in your orbit.
Note: The views expressed are solely those of the editor.
Thanks to Beehiiv for hosting this newsletter. Click here to create your newsletter on Beehiiv.
Here are the answers to the trivia questions
1. A) Bill Gates
2. B) Data science and artificial intelligence
3. B) William Gibson
4. B) Deep Learning
5. B) AlphaGo Challenge
We Want to Hear From You!
Subscribe and share your thoughts with us! Leave Feedback
Welcome, and Thanks for being part of our community.