AI Safety Tests Reveal Deceptive AI Responses
The UK AI Safety Institute had identified unprecedented malicious behaviour in AI models from Anthropic and OpenAI during safety testing.
Category
Product launches, AI, and software trends.
The UK AI Safety Institute had identified unprecedented malicious behaviour in AI models from Anthropic and OpenAI during safety testing.
A new report from J.P. Morgan Asset & Wealth Management warns that artificial intelligence has accelerated cyber threats, reducing the average window between software vulnerabilit...
The Netherlands had builded the world's first power plant generating electricity by mixing salt water with fresh water, showcasing a novel renewable energy method.
Engineers at Texas A&M University are creating microscopic, fish-shaped robots that use light-activated particles to extract lithium from seawater, offering a sustainable energy r...
IAF Chief Air Chief Marshal Amar Preet Singh advocates for 100% domestic drone production to remove foreign supply risks and highlights the new Directorate of Aerospace Design’s r...
India's DigiLocker initiative expands internationally as work begins in Kenya and Cuba, marking a significant step in global digital identity services using India Stack technology.
Goldman Sachs forecasts Generative AI will increase India's annual labor productivity by approximately 0.4% over the next decade, mainly enhancing roles in knowledge sectors.
Himachal Pradesh has introduced AI tools named 'WhereToPlant' and 'WhatToPlant' to improve the accuracy and success rate of forest restoration efforts.
Satya Nadella cautions that companies depending on a single AI model without their own infrastructure risk failure in the evolving AI landscape.