Microsoft CEO Satya Nadella is using 'Super Intelligence' to describe language models, a term favored by former President Trump, in a recent blog post. This shift marks a departure from the standard scientific terminology for artificial intelligence, as Nadella aligns with Trump's preferred language.
Nadella's post targets OpenAI and Anthropic, suggesting they pose risks akin to insider security threats. He argues that AI models should be treated with caution due to their potential to make mistakes or be compromised, despite their access to sensitive data and control over critical systems.
The CEO proposes measures such as model diversity, logging all actions, independent auditing, and human oversight to ensure safety. These steps aim to reduce reliance on a single model and increase transparency, according to Nadella's blog.
"Not because they are necessarily malicious, but because any sufficiently capable actor with access to important systems can make mistakes or be compromised," said Nadella. He emphasizes the need for a containment and control architecture to manage these risks effectively.
The move comes amid Microsoft's broader strategy to position itself as an open AI ecosystem, while quietly replacing OpenAI and Anthropic models with cheaper in-house alternatives. This reflects a business-driven approach to maintain margins in a competitive market.
Microsoft did not specify how it plans to implement these safety measures, and the effectiveness of Nadella's arguments remains uncertain. The company's focus appears to be on managing its business strategy amid rapid AI advancements.
Source: thedecoder
