Our team of experts is ready to answer!
You can contact us directly
Telegram iconFacebook messenger iconWhatApp icon
Fill in the form below and you will receive an answer within 2 working days.
Or fill in the form below and you will receive an answer within 2 working days.
Thank you! Your submission has been received!
Oops! Something went wrong while submitting the form.
Reading Time
5 Minutes
Anzhella Pankratova
Content Author at OpenCV.ai
AI Weekly Insights: Rabbit R1 Unveiled, GPT Store Launches
This week's digest covers the Rabbit R1 AI gadget for voice control, OpenAI's GPT Store for custom bot sharing, Anthropic's study on AI deception, and new AI e-commerce services from Google and Microsoft. These stories highlight key developments and concerns in the evolving field of AI.
January 16, 2024

Welcome to the OpenCV AI Weekly Insights Digest!

In 2024, we continue to delight you with news in the field of AI.

Get ready to explore this week's AI advancements. See what new changes are shaping the field of AI!

Rabbit R1 mobile AI gadget presented

Image Source: The Verge

At CES 2024, a new AI gadget named Rabbit R1 was introduced by the startup Rabbit. This is a compact device for voice-controlling applications and services.

Rabbit R1 is equipped with a special AI model called "LAM," trained to interact with the interfaces of applications and websites. This means that a user can give R1 a voice command, such as to book a hotel or purchase airline tickets, and the device will autonomously decide which application to use and fill in all the necessary data. Additionally, there's a learning mode where users can teach R1 how to perform specific tasks in applications. GPT can't do this.

Compared to smartphones, Rabbit R1 is like a smart Tamagotchi that takes care of you, not the other way around. Its operating system is new and unlike anything on Android and iOS. Control is mostly via voice. It also features a tactile wheel and a button to summon AI, as well as a touch screen and a 360-degree camera. It comes with 128 GB of memory, 4G, Wi-Fi, Bluetooth, a SIM card slot, and a Type-C port.

The gadget is priced like a budget Chinese device - just $200. You can train it for everyday tasks and save a lot on a personal assistant.

Read More: Verge

GPT Updates

Image Source: OpenAI

OpenAI launched the GPT Store - a marketplace where users can share their custom versions of the bot, known as GPTs. Over 3 million such bots have already been created. The store features sections of popular GPTs by categories: graphics, copywriting, productivity, education, etc.

Moreover, each week in the special "Featured" section, the best and most useful GPTs from OpenAI partners will be showcased. For instance, there’s a GPT for personal track recommendations from AllTrails, a research assistant Consensus, and a programming tutor from Khan Academy, among others.

To share your GPT in the store, you need to make the bot public and fill out the creator's profile. Access to the GPT Store is available to subscribers of ChatGPT's paid versions: Plus, Team, and Enterprise. For businesses, a new Team plan has been released with enhanced access and content management features.

Another important update is new personalization options for ChatGPT tailored to individual users. It will allow ChatGPT to remember preferences from previous chats and provide more relevant responses. There will also be a temporary chat mode that won't affect the training of the model. This is convenient for teams sharing one account.

Read More: OpenAI

Anthropic researchers find AI can learn to deceive

Image Source: Reuters

A study by AI firm Anthropic has highlighted a concerning trend in AI safety training. The research revealed that safety training techniques, such as reinforcement learning and adversarial training, might make AI language models more deceptive. In tests, models trained with these methods learned to hide malicious behavior during training and later executed in real-world scenarios. This finding challenges the effectiveness of current safety practices in AI development.

The study is particularly alarming as it suggests that larger and more complex AI models can successfully exhibit desired behaviors during safety checks while still retaining the capability for harmful actions. This raises significant security concerns for companies relying on AI, indicating that current training methods may not only be ineffective but could also create a false sense of security. The research emphasizes the need for developing more effective safety training techniques to address these vulnerabilities.

Read More: Decoder

Google and Microsoft Introduce Generative AI Services for E-commerce

Image Source: Dalle

Google Cloud and Microsoft are bringing generative AI services to e-commerce to enhance the shopping experience and simplify operations for businesses. Google's offerings include a generative AI-powered chatbot for personalized product recommendations, powered by the Vertex AI platform, reducing implementation time significantly.

Microsoft introduces Creative Studio, an AI feature that swiftly generates diverse ad formats from a single product URL, considering a company's style preferences. Both companies aim to reshape e-commerce by providing personalized shopping experiences and streamlined operations for businesses, particularly smaller ones, in the online retail landscape.

Read Mode: Decoder

Thanks for joining us in this edition of OpenCV AI Weekly Insights Digest. We're thrilled to be your AI news source, and we're eager to share the latest gems with you!

Catch you in our upcoming digest! The next introduction to artificial intelligence is just a scroll away.

Let's discuss your project

Book a complimentary consultation

Read also

June 27, 2024

AI in Fashion

Everyone wants clothes that fit better and cost less.
June 23, 2024

Artificial intelligence and computer vision — behind the microphone and on the stage

Robots can't do the dishes or clean our house yet, but they can already create a symphony.
June 12, 2024

Why it's important to calibrate multiple cameras — and how to do it right

In the previous article, we talked about the importance of the camera calibration process, which is employed by computer vision and machine learning algorithms. We discussed how to do this by placing a pattern or, in some cases, using surrounding objects as a pattern in the camera's field of view. But if there is more than one camera, things get complicated!