Welcome to Get The Gist, where every weekday we share an easy-to-read summary of the latest and greatest developments in AI—news, innovations, and trends—all delivered in under 5 minutes! ⏱

In today’s edition:

• Microsoft Gives Free Access to OpenAI’s o1 Model in Copilot

• Google Rolls Out Gemini 2.0 Pro Experimental

• Meta Enhances AI Labels For Ads on Facebook & Instagram

• OpenAI Files Trademark for Humanoid Robots & AI-Powered VR Headsets

• And more AI news….

Top Developments

Microsoft

Microsoft Gives Free Access to OpenAI’s o1 Model in Copilot

The Gist: Microsoft is making OpenAI’s o1 model, previously available only to ChatGPT Pro users for $200 per month, free for all Copilot users. This advanced AI model enhances reasoning and problem-solving in science, coding, and math.

Key Details

• The o1 model is accessible through Copilot’s "Think Deeper" feature, which can be toggled on for more in-depth responses.

• While responses take longer (about 30 seconds), Microsoft says this allows for deeper analysis and better answers.

• Think Deeper helps users with career advice, step-by-step project guides, and detailed topic explorations.

• The feature does not access the internet, and its knowledge is current up to October 2023, though improvements are in development.

Google

Google Rolls Out Gemini 2.0 Pro Experimental

The Gist: Google has introduced its latest AI model, Gemini 2.0 Pro Experimental, with little fanfare, revealing it only in a changelog for its Gemini chatbot app. The model promises improved factual accuracy and stronger coding and math capabilities, but remains in an early preview phase.

Key Details:

• Gemini 2.0 Pro Experimental is available to Gemini Advanced users through Google One AI Premium and Workspace add-ons.

• The model enhances coding and mathematical problem-solving but may exhibit unexpected behaviors.

• Unlike other Gemini models, it lacks real-time internet access and is incompatible with some app features.

• Google also made its Gemini 2.0 Flash model, launched in December, available to all Gemini app users.

Meta

Meta Enhances AI Labels For Ads on Facebook & Instagram

The Gist: Meta is expanding its AI labeling system to make it clearer when ads on Facebook and Instagram contain AI-generated or heavily edited content. This move aims to increase transparency as AI-generated media becomes more prevalent.

Key Details:

• Meta will now label ads created with third-party AI tools like OpenAI and Google, in addition to its own AI-generated content.

• Labels will appear in different locations based on the content, with more prominent placement for AI-generated photorealistic humans.

• Minor AI edits without significant alterations or photorealistic elements may not receive a label.

• The update comes as Meta phases out its fact-checking programs, making clear AI disclosures more important.

Quick Gist

Adobe introduced new generative AI features in Acrobat, enabling users to better interact with contracts through contract intelligence, summaries, and version checking tools (Read More).

China's DeepSeek.com surged to become the second most popular AI chatbot, with traffic reaching 49 million visits in a week, surpassing Google's Gemini (Read More).

Google is integrating its AI model Gemini into Google Maps, allowing users to ask questions about locations while exploring the map (Read More).

The European Union's AI Act is now in effect, banning high-risk AI systems and imposing substantial fines for non-compliance (Read More).

OpenAI has filed a trademark application for humanoid robots and AI-powered VR headsets, signaling a potential shift towards hardware development (Read More).

Samsung Electronics chief Jay Y. Lee met with OpenAI CEO Sam Altman and SoftBank's Masayoshi Son to explore potential collaboration on OpenAI's $500 billion Project Stargate (Read More).

Anthropic introduced "constitutional classifiers" for its Claude 3.5 model, reducing jailbreak success rates from 86% to just 4.4% (Read More).

Alibaba launched an upgraded AI model, Qwen 2.5-Max, claiming it surpasses competitors including GPT-4o and DeepSeek-V3 (Read More).

Realbotix is set to enhance its humanoid robots with large language model integration and advanced customization features (Read More).

Google's Pixel 9a will not include the Gemini Advanced perks offered with other Pixel devices (Read More).

Australia's federal government has banned the use of the Chinese AI model DeepSeek on government devices due to national security concerns (Read More).

THAT’S IT FOR TODAY!