Artificial Intelligence: News, Business, Science

AI in practice

Apr 21, 2024Apr 21, 2024

Hundreds of examples in prompts can significantly boost LLM performance, study finds

AI research

Apr 18, 2024Apr 18, 2024

US Air Force successfully tests AI-controlled fighter jets in simulated dogfights

AI and society

Apr 15, 2024Apr 15, 2024

To avoid AI-driven "knowledge collapse", humans must actively preserve specialized expertise

AI research AI and society AI in practice

AI research

Apr 24, 2024Apr 24, 2024

Short

According to a recent survey conducted by ZoomRx, two-thirds of the 20 largest pharmaceutical companies prohibit their employees from using ChatGPT due to security concerns regarding sensitive internal data. Nevertheless, many life sciences professionals use ChatGPT regularly, even though 83 percent say the technology is "overrated." Despite the potential for increased efficiency and effectiveness through the use of AI in drug development, the majority of pharmaceutical companies are cautious about the technology due to security risks. Most companies see AI primarily as a means to reduce costs, while concerns about data security and privacy remain. But there are exceptions. According to OpenAI, Moderna is using ChatGPT in a pilot project called Dose ID to analyze and visualize large data sets and determine the optimal vaccine dose.

Fierce Pharma

AI and society

Apr 24, 2024Apr 24, 2024

Short

Japan is considering creating a new dialogue forum with like-minded nations to discuss global regulations for the responsible use of generative AI technology. Prime Minister Fumio Kishida is expected to make an announcement at an OECD ministerial meeting in Paris in early May, according to Mainichi's sources. Japan aims to build broader support for the "Hiroshima AI Process," a G7 initiative to discuss global rules for AI. As generative AI tools such as ChatGPT rapidly advance, there is growing concern that misinformation spread by AI misuse could threaten democracy and political stability. By taking a leading role in international regulation of AI, Japan hopes to promote both the development and oversight of the technology. The non-binding guidelines agreed at the G7 summit in Hiroshima aim to promote the creation of safe and trustworthy AI systems.

Mainichi

AI research

Apr 24, 2024Apr 24, 2024

Update

THE DECODER

Hundreds of examples in prompts can significantly boost LLM performance, study finds

US Air Force successfully tests AI-controlled fighter jets in simulated dogfights

To avoid AI-driven "knowledge collapse", humans must actively preserve specialized expertise

Adobe's VideoGigaGAN turns low-res videos into high-res, detailed clips

Current LLMs "undertrained by a factor of maybe 100-1000X or more" says OpenAI co-founder

OpenAI's new 'instruction hierarchy' could make AI models harder to fool

Microsoft's small and efficient LLM Phi-3 beats Meta's Llama 3 and free ChatGPT in benchmarks

AI search engine Perplexity.ai hits billion dollar valuation and releases new enterprise offering

A simple trick makes Meta's Llama 3 model go rogue

Adobe unveils Firefly Image 3 and a major upgrade to the AI capabilities of Photoshop

Nick Bostrom sparked a global AI debate, but his institute couldn't survive academia

Hundreds of examples in prompts can significantly boost LLM performance, study finds

US Air Force successfully tests AI-controlled fighter jets in simulated dogfights

To avoid AI-driven "knowledge collapse", humans must actively preserve specialized expertise

Bank details