Welcome to Daily Zaps, your regularly-scheduled dose of AI news ⚡️

Here’s what we got for ya today:

• 🥇 AI gymnastics judging system

• 🔍 SearchGPT

• 🇺🇸 NIST draft guidance from U.S. AI Safety Institute

• 💸 OpenAI could lose $5 billion this year

Let’s get right into it!

Sign Up | Advertise| Tools

BIG TECH

AI gymnastics judging system

The International Gymnastics Federation (FIG) has introduced an AI-assisted judging system developed with Fujitsu to enhance the accuracy and fairness of gymnastics scoring. This technology, known as the Judging Support System (JSS), uses advanced cameras and 3D modeling to provide a second opinion on the difficulty scores of gymnasts' routines. Initially intended to be ready for the 2021 Tokyo Olympics, it finally debuted at the 2023 World Championships in Antwerp. While the system aims to address issues of human error and bias in scoring, its implementation is costly and complex, raising questions about its necessity and potential applications beyond gymnastics, such as in healthcare and theft prevention.

Despite these advancements, the JSS is currently used primarily for validating human judges' decisions rather than replacing them. This technology marks a significant step towards more objective scoring, but it also highlights the challenges of integrating AI into sports where subjective judgment plays a crucial role. The hope is that the JSS will eventually provide more consistency and transparency in judging, helping to ensure that gymnasts' performances are evaluated as fairly and accurately as possible.

STARTUPS

SearchGPT

OpenAI has announced its entry into the search market with SearchGPT, an AI-powered search engine that organizes and makes sense of information from the internet in real-time. Unlike traditional search engines, SearchGPT provides summarized findings with attribution links and allows users to ask follow-up questions or explore related links. The service, developed in collaboration with news partners, emphasizes clear attribution and links to sources. As a prototype, it allows OpenAI to refine the product while addressing potential issues.

Initially available to 10,000 test users, SearchGPT is powered by the GPT-4 model and aims to integrate search features into ChatGPT eventually. The launch marks a significant challenge to Google's dominance in the search market and highlights OpenAI's growing influence and operational costs.

GOVERNMENT

NIST draft guidance from U.S. AI Safety Institute

The Department of Commerce’s National Institute of Standards and Technology (NIST) has released three final guidance documents and a software package as part of implementing the October 2023 executive order on artificial intelligence. These documents aim to help developers manage generative AI risks and align with NIST’s Secure Software Development and AI Risk Management Frameworks. Additionally, the U.S.

AI Safety Institute has issued draft guidance on mitigating misuse risks for dual-use foundation models, with public comments due by September 9. The software package assists AI developers in assessing the impact of adversarial attacks on AI systems. Commerce Secretary Gina Raimondo emphasized the importance of these tools in ensuring the safe development and deployment of AI.

STARTUPS

OpenAI could lose $5 billion this year

OpenAI is one of the fastest-growing businesses in history but also one of the costliest to operate. The company could lose up to $5 billion this year, according to analysis based on internal financial data. OpenAI's expenses include $4 billion for renting Microsoft's servers, $3 billion for AI training costs, and $1.5 billion for staffing.

Despite generating revenue from various sources, including $2 billion annually from ChatGPT and over $80 million per month from API access, OpenAI's potential revenue of up to $4.5 billion falls short of its $8.5 billion in operating costs. This financial strain may necessitate raising additional funds soon. While OpenAI benefits from discounted rates for Microsoft's servers, reducing costs and increasing revenue through new products and improved models remain critical for future sustainability.