OpenAI assembles a team for "catastrophic risks"

Pivot 5: 5 stories. 5 minutes a day. 5 days a week.

1. OpenAI assembles a team for "catastrophic risks"

OpenAI has established a dedicated team to address the "catastrophic risks" linked to AI. The team's mandate is to "track, evaluate, forecast, and protect" against significant AI-induced challenges, encompassing nuclear threats and other hazards like "chemical, biological, and radiological threats." The team will also focus on "autonomous replication", where AI self-replicates, and other risks such as AI's capability to deceive humans and cybersecurity vulnerabilities.

Aleksander Madry, taking a hiatus from his directorial role at MIT’s Center for Deployable Machine Learning, will spearhead this initiative. OpenAI emphasizes that the team will formulate a "risk-informed development policy" detailing the firm's efforts in evaluating and overseeing AI models.

Read the full story here 

2. Google Maps: AI refines search and navigation

Google is introducing several updates to Maps, incorporating AI-powered features aimed at simplifying user searches and exploration. One notable enhancement allows users to search for specific items, such as “animal latte art”, and receive photo results based on AI and advanced image recognition. These images, sourced from user-shared content on Google Maps, facilitate the discovery of new locations that match the user's search criteria.

Additionally, the Maps navigation interface will soon depict surroundings more authentically, showcasing realistic buildings and improved lane details, enhancing the user's orientation during drives.

Read the full story here 

3. Forbes introduces AI search tool: Adelaide

Forbes has unveiled a beta version of its generative AI search platform, Adelaide, developed in collaboration with Google Cloud. Adelaide offers readers personalized search experiences. Users can pose specific questions or general topics and receive article recommendations and summarized answers, provided the content falls within Forbes' purview. The AI system has been trained on Forbes articles from the past year, encompassing news and list coverage.

By tapping the Adelaide button on Forbes' website, readers can inquire, for instance, about Nebraska's wealthiest individual. The system would then generate a summary about Warren Buffett and suggest related articles. Supitskiy envisions expanding Adelaide's knowledge base to encompass Forbes' entire archive, tracing back to its inception in 1917.

Read the full story here

4. Shutterstock's AI Offers Customizable Stock in 700M Image Library

Shutterstock, Inc. has introduced innovative AI-powered editing features, offering limitless customization options for its vast library of over 700 million stock images. Currently in beta, these AI features utilize OpenAI technology, enabling users to generate or modify any image within Shutterstock's collection, streamlining content ideation and production.

The suite, set to be demoed on November 9, includes capabilities like the "Magic Brush" for image modifications and "AI Image Generator" for creating visuals based on descriptions. Additionally, Shutterstock ensures artists' compensation for edited images and upholds ethical AI practices in the creative sector.

Read the full story here 

5. AI sound recordings monitor forest recovery

In a recent study in Ecuador, scientists are leveraging sound recordings and AI to assess biodiversity recovery in tropical forests. Acoustic monitoring combined with AI tools provides insights into the regenerating forest plots' health and biodiversity. This innovative approach offers a promising avenue for understanding and preserving the essential ecosystems of tropical forests.

The World Economic Forum emphasizes the significance of such advancements, highlighting the potential of technology in aiding conservation efforts and ensuring the sustainability of vital ecosystems.

Read the full story here