- The Daily Update
- Posts
- Top AI employees warn about risks of advanced AI
Top AI employees warn about risks of advanced AI
ALSO 'Xi Jinping LLM' and Elon Musk's latest move
Jack Shields
June 05, 2024
Welcome back!
It’s officially been a year since I started writing The Daily Update. Thank you for reading this newsletter everyday and being a part of our AI community. Let’s get into today’s news.
In today’s Daily Update:
🗞️ Top AI employees issue warning on risks of advanced AI
🤖 China tests ‘Xi Jinping LLM’
📸 Elon Musk diverts thousands of Nvidia chips from Tesla to X
🚨 AI Roundup: Four quick hits
Read time: 2.5 minutes
TOP STORY
🗞️ Top AI employees issue warning on risks of advanced AI
DALL-E 3
Yesterday a group of current and former employees at top AI companies published an open letter warning that AI will threaten humanity without increased oversight.
What you should know:
The letter was signed by 13 employees from OpenAI, Anthropic and Google DeepMind.
According to the letter, risks include worsening existing inequalities, increasing misinformation and losing control of autonomous AI systems that could wipe out humanity.
The group says loosely regulated AI companies are prioritizing profits over safety as they pursue AGI.
The employees want their employers to waive nondisclosure agreements and allow workers to anonymously raise concerns.
The big picture: OpenAI is at the center of this debate after being repeatedly criticized for sacrificing safety protocols to launch new products, some of which have already been weaponized by foreign adversaries. Because there is little government oversight of AI companies, employees are currently the only people who can hold industry giants accountable. Regardless of your take on the issue, this letter raises several valid concerns about the state of the AI sector today.
AI INSIGHT
🤖 China tests ‘Xi Jinping LLM’
DALL-E 3
The China Cyberspace Research Institute has reportedly developed a language model based directly on President Xi Jinping’s philosophy.
Key points:
The model is designed to accurately reflect the values of the Communist Party of China.
It functions exactly like other LLMs, able to answer questions, write new content and translate between languages.
Unlike other systems, the model was trained on a limited dataset that only includes pieces of information consistent with the Chinese Communist Party’s values.
The China Cyberspace Research Institute is currently testing the model internally and has not released it to the public.
Food for thought: Harsh restrictions on training data can negatively impact the performance of LLMs, meaning China’s restrictive policy could actually help the U.S. in the AI race. At a broader level, politically aligned LLMs create the exact types of risks that the employees in today’s top story want to avoid.
BUSINESS SPOTLIGHT
📸 Elon Musk diverts thousands of Nvidia chips from Tesla to X
DALL-E 3
Elon Musk reportedly ordered Nvidia to divert thousands of AI chips reserved for Tesla to his social media company X.
The details:
Musk told shareholders in April that Tesla would increase its arsenal of Nvidia H100s from 35,000 to 85,000 by the end of this year.
Internal emails obtained from Nvidia employees suggest that Musk redirected many of the chips to X and its subsidiary xAI.
The move could delay Tesla’s acquisition of $500 million worth of Nvidia processors by several months.
Musk said on X that Tesla was unable to accept Nvidia’s GPUs because the company’s factory in Austin, Texas is still incomplete.
Why it matters: This could mark a shift away from self-driving technology as Musk tries to build a competitor to OpenAI’s ChatGPT. Apple made a similar move earlier this year when it canceled its years-long electric vehicle project to focus on generative AI. Regardless, Tesla shares dropped around 1% yesterday as the company continues to struggle in the EV market.
MORE TRENDING NEWS
🚨 AI Roundup: Four quick hits
DALL-E 3
Sam Altman says OpenAI has enough data to train the next generation of AI.
ChatGPT, Claude and Perplexity experience outages simultaneously. The cause remains unknown.
Microsoft reaches multibillion-dollar AI-focused partnership with Tokyo-based Hitachi.
Stability AI plans to launch Stable Diffusion 3 “medium” image generator for public use on June 12.
THAT’S ALL FOR TODAY