Google dropped its promise on using AI for weapons

Hello everyone! Here’s what you need to know about AI today:

👉 Google drops pledge not to use AI for weapons

👉 Boston Dynamics joins forces with its former CEO to build better robots

👉 Google has officially launched Gemini 2.0 Pro

and many more!

📧 Did someone forward you this email? Subscribe here for free to get the latest AI news everyday!

Read time: 5 minutes

What’s going on: Google has made a significant policy shift by removing its long-standing prohibition against using AI for weapons and surveillance systems. This change, which was quietly implemented, marks a departure from the ethical guidelines Google had established in 2018, known as Google's AI Principles. (Read the original principles)

What does it mean: The removal of these prohibitions has sparked a debate within the tech community about the implications for AI safety and ethical development. The updated principles now focus on mitigating unintended or harmful outcomes rather than outright banning certain uses.

More details: 

• Critics like Tracy Pizzo Frey, a former Google employee who was involved in implementing these principles, have expressed concerns over this shift.

• The alteration in Google's AI policy could also change how other tech companies approach ethical AI development.

• This change might lead to a focus on rapid AI development at the expense of ethical considerations, as AI moves into more sensitive areas like autonomous weapon systems.

What’s going on: Boston Dynamics has announced a partnership with the Robotics & AI Institute (RAI) to advance the reinforcement learning capabilities of its Atlas humanoid robot. Founded by Marc Raibert, who previously led Boston Dynamics, the Institute will collaborate to improve how Atlas learns new tasks and interacts with its environment.

What does it mean: This collaboration is specifically aimed at improving simulation-based learning for Atlas. By refining how the robot learns within a digital environment, the transition to real-world applications can become more seamless and effective. This strategic move by Boston Dynamics could potentially set new benchmarks in robotics by accelerating the pace at which robots like Atlas can adapt and learn from their surroundings.

More details:

• The focus will be on dynamic movements, whole body loco-manipulation, and full-body contact strategies, skills essential for the robot to function in varied and unpredictable environments.

• This partnership is part of a broader trend where robotics companies are increasingly developing their own AI models, much like Figure AI, to tailor solutions more precisely to their hardware.

What’s going on: Last week, Google accidentally announced the release of its Gemini 2.0 Pro through an update in their changelog and quickly removed the update afterwards. But yesterday, they officially launched an experimental version of Gemini 2.0 Pro in Google AI Studio, Vertex AI and in the Gemini app for Gemini Advanced users. Google claims this models is their best model yet for coding performance and complex prompts.

What does it mean: This model makes advanced AI features available to more people by being cheaper, which could bring in more developers. Google is launching these models to stay competitive in the fast-changing world of AI, where they face rivals like DeepSeek-R1 and OpenAI’s o1 and o3-mini models.

More details: 

• The context window of Gemini 2.0 Pro is 2 million tokens, meaning it can take up to 1.5 million words at once. It means you can feed it 15 to 20 average-sized novels and it can easily process it!

• Gemini 2.0 Pro has access to Google Search, can execute code on behalf of users, and supports function calling and structured output.

• You can easily access this flagship model on Google AI Studio for free. From the “Model” dropdown, select Gemini 2.0 Pro Experimental 02-05 and enjoy!

• Additionally, Google has released Gemini 2.0 Flash-Lite, which is the most cost-efficient model in the Gemini lineup.

👋 Do you enjoy reading so far? Be an AI ambassador by sharing the link below with your friends. You can also forward this email to them! Thanks!
the79newsletter.com/subscribe

⭐ OpenAI just announced that “ChatGPT Search” is now available to everyone on chatgpt.com and no sign up is required.

🤖 Amazon is reportedly preparing to unveil a next-generation, AI-powered version of Alexa at an event on February 26, which can handle multiple commands in sequence.

💼 Workday, a leading enterprise HR platform, announced it is laying off nearly 2,000 employees, or about 8.5% of its workforce, to refocus investments towards AI and global expansion.

🤯 Researchers at Stanford and University of Washington have developed an open-source alternative to OpenAI's o1 reasoning model for less than $50 (read the paper they have published).

🔍 LinkedIn is testing an AI tool designed to enhance job searches by uncovering hidden job opportunities that traditional search queries might miss.

⚛ Google's head of Quantum AI, Hartmut Neven, predicts that commercial quantum computing applications will be out within 5 years.

AI + business plan

Scale AI - Senior Forward Deployed Engineer, GenAI

Rocket Money - Senior ML Engineer, Product

Substack - Machine Learning Engineer

Garner Health - Data Scientist

Thank you for staying with us like always! If you are not subscribed, subscribe here for free to get more of these emails in your inbox! Cheers!