Copilot+ PCs

Today:

• Copilot+ PCs

• Google Expands Finnish Data Center

• OpenAI Suspends Controversial Voice

• Inflection AI EQ-Powered Bots

• UK, South Korea Lead AI Talks

• AI Beats Humans in Mind Tests

Introducing Copilot+ PCs

Microsoft introduced Copilot+ PCs, a new category of Windows PCs designed for AI, at a special event. These PCs feature powerful new processors, all-day battery life, and advanced AI models. Key features include Recall for easily finding files, Cocreator for AI-powered image creation, and Live Captions for translating audio. 

Available from Microsoft Surface and partners like Acer, ASUS, Dell, HP, Lenovo, and Samsung, pre-orders start today at $999, with availability beginning June 18. Copilot+ PCs promise superior performance, efficiency, and security, marking a significant evolution in PC design centered around AI innovation.

MICROSOFT

Google invests 1 billion euros in Finnish data centre to drive AI growth

Google announced a €1 billion investment to expand its data center in Finland to support AI growth. The cooler climate, tax breaks, and renewable energy availability make the Nordic region ideal for data centers. Finland’s wind power capacity has surged, making renewable energy available even for energy-intensive data centers. 

Google's Hamina center, operating with 97% carbon-free energy, will also contribute heat to local infrastructure. This investment aligns with Google's goal to achieve net zero emissions by 2030. Additionally, new data centers are planned in the Netherlands and Belgium to further support AI advancements.

REUTERS

ChatGPT suspends AI voice that sounds like Scarlett Johansson

OpenAI has withdrawn the "Sky" voice option from ChatGPT after it sparked widespread comparison to Scarlett Johansson's character in the film Her. The voice, introduced at the debut of the new ChatGPT-4o model, was noted for its personable and responsive nature. 

Despite OpenAI CEO Sam Altman's suggestive tweet and media comparisons, the company clarified that "Sky" was not based on Johansson and was voiced by a different actress. OpenAI emphasized its respect for the voice acting industry and cited safety concerns following the resignation of key team members. The incident highlights ongoing tensions over AI’s impact on the entertainment industry.

THE GUARDIAN

Inflection AI reveals new team and plan to embed emotional AI in business bots

Inflection AI, after the departure of Mustafa Suleyman to Microsoft, introduced a new leadership team including CEO Sean White, CTO Vibhu Mittal, COO Ted Shelton, and Head of Product Ian McCarthy. Despite the exodus, co-founder Reid Hoffman continues to back the company, which has significant funding to focus on emotional intelligence (EQ) in AI. 

Inflection AI's mission is to create empathetic chatbots for businesses, leveraging a strong foundation in EQ to offer personalized and considerate interactions. The company plans to build customized AI assistants for businesses, enhancing customer support and internal employee interactions.

VENTUREBEAT

South Korea, UK to host AI summit in Seoul as risks mount

South Korea and the UK will host a global AI summit in Seoul this week to address the rapid advancements and associated risks of AI. British Prime Minister Rishi Sunak and South Korean President Yoon Suk Yeol will lead the virtual event, focusing on AI safety, innovation, and inclusion. 

The summit aims to fill gaps in global AI governance and will feature leaders from the G7, Singapore, Australia, and China. The discussions will cover the extensive risks posed by AI, including labor market impacts, hacking, and environmental concerns. An agreement on AI governance is expected to be adopted.

REUTERS

AI models can outperform humans in tests to identify mental states

New research published in Nature Human Behavior reveals that some large language models (LLMs) can outperform humans in tests designed to measure "theory of mind," the ability to understand others' mental states. These AI models, including OpenAI's GPT-4 and Meta’s Llama, showed impressive results in tasks such as recognizing irony and understanding indirect comments. 

However, they still lack true human-like comprehension and empathy. The study underscores the potential and limitations of AI in mimicking human social intelligence, highlighting the need for careful evaluation of what these models can genuinely achieve in terms of emotional and social understanding.

MIT