AI News

Follow the latest AI and automation news

22 Sept 202611 Rabiʻ II 1448 AH

Parallel cut research time and cost in half with GPT‑6 Astra

Parallel has launched the GPT-6 Astra model, enabling its agents to research and synthesize labor market data more efficiently. The new model can complete tasks in half the time and cost compared to previous models, showcasing a significant advancement in AI technologies. This development enhances companies' ability to make decisions based on accurate and timely data, positively impacting business strategies in the market.

Governance

22 Sept 202611 Rabiʻ II 1448 AH

Meta patches Muse exploit that let attackers control the AI agent

Meta has issued a patch for its Muse macOS app following the discovery of a zero-day vulnerability. This flaw, identified by security researcher Patrick Wardle, allowed potential attackers to take control of Muse accounts by exploiting an undocumented setting within the app. The vulnerability enabled attackers with local access to redirect transcription processing from Meta's servers to their own endpoint. This means that anyone with local access to the user's device could potentially access Muse accounts, posing a significant privacy risk.

Models

22 Sept 202611 Rabiʻ II 1448 AH

Xiaomi's affordable flagship AI leads the open models, and Anthropic says Claude helped get it there

Xiaomi has announced its MiMo-V2.6-Pro model, which leads the open AI models market with competitive pricing. This success is attributed to the company's investment in reinforcement learning, which cost $2.62 million, allowing it to offer an effective model at a low price. However, Xiaomi faces accusations from Anthropic of siphoning training data from the Claude model, raising ethical questions about data usage in developing intelligent models.

💡

AutoScheduler launches warehouse app builder for logistics teams

How might this tool impact warehouse management efficiency?

Business

Warnings About Risks of AI Agents in Online Shopping

22 Sept 202611 Rabiʻ II 1448 AH

Business

22 Sept 202611 Rabiʻ II 1448 AH

Warnings About Risks of AI Agents in Online Shopping

A group of banks, including NatWest and Bank of America, warned that the use of AI agents in online shopping could increase risks of fraud and data privacy violations. Companies like OpenAI and Google are promoting this technology, raising concerns about consumer protection. Data from John Lewis indicated a rise in AI agent searches to 2.5%, reflecting growing reliance on this technology.

Models

OpenAI says its internal model solved over 100 long-standing math problems after just a month of training

22 Sept 202611 Rabiʻ II 1448 AH

Models

22 Sept 202611 Rabiʻ II 1448 AH

OpenAI says its internal model solved over 100 long-standing math problems after just a month of training

OpenAI announced that its internal model solved over 100 math problems after just one month of training. The company is facing criticism from mathematicians, prompting it to support an independent advisory group at the Institute for Advanced Study. However, OpenAI has excluded its research pace from the group's advisory role. This development highlights the challenges OpenAI faces in the mathematics domain and underscores the importance of transparency in AI research. It may influence how mathematical models are evaluated in the future.

Models

Don’t be fooled by this summer of AI hype

22 Sept 202611 Rabiʻ II 1448 AH

Models

22 Sept 202611 Rabiʻ II 1448 AH

Don’t be fooled by this summer of AI hype

Recent months have seen a surge of hype around AI, with Anthropic claiming its model Claude Mythos excels at finding software vulnerabilities. This was followed by the OpenAI-Hugging Face hacking incident, leading to disclosures from both Anthropic and Meta regarding similar issues. Cybersecurity experts suggest that the narrative is more about OpenAI's negligence in adopting basic security practices than rogue AI models.

Business

SoftBank Secures Over $20 Billion for OpenAI Investments

22 Sept 202611 Rabiʻ II 1448 AH

Business

22 Sept 202611 Rabiʻ II 1448 AH

SoftBank Secures Over $20 Billion for OpenAI Investments

SoftBank attracted over $20 billion in initial demand for high-risk bonds aimed at funding the group's investments in OpenAI. The company seeks to raise more than $11 billion, including a $10 billion dollar tranche and a €1 billion euro tranche. Initial demand indicates high yields of around 10% on longer-term dollar bonds. This bond deal comes as investors test their willingness to fund rapid expansion in the AI sector, amid concerns about spending levels and associated risks. Led by Masayoshi Son, SoftBank has committed approximately $65 billion to invest in OpenAI, positioning itself as one of the largest investors in this field.

Arab AI Scene

NVIDIA Confirms Egypt's Potential as a Regional AI Development Hub

22 Sept 202611 Rabiʻ II 1448 AH

Arab AI Scene

22 Sept 202611 Rabiʻ II 1448 AH

NVIDIA Confirms Egypt's Potential as a Regional AI Development Hub

NVIDIA has confirmed that Egypt has the necessary components to become a regional hub for AI application development, highlighting the current phase as an opportunity to accelerate the local ecosystem. Ulu Guillemin, NVIDIA's VP for Europe, the Middle East, and Africa, stated that the world is at a pivotal moment in technology, with direct implications for individual lifestyles and economic development. Guillemin explained that his visit aims to explore Egypt's plans for AI deployment, asserting that the timing is right for this initiative, focusing on infrastructure development and enhancing local innovation.

Governance

A New Tool Found Malware That’s Guided by an AI Hive Mind—No Humans in Sight

22 Sept 202611 Rabiʻ II 1448 AH

Governance

22 Sept 202611 Rabiʻ II 1448 AH

A New Tool Found Malware That’s Guided by an AI Hive Mind—No Humans in Sight

Cisco Talos researchers have developed a new framework for identifying malware and hacking tools that utilize AI chatbots. This framework enables rapid detection of potential threats that may be employed by criminal groups. Research has shown that these tools can be more complex than previously anticipated, necessitating the development of new strategies to combat them.