🚨 OpenAI New Product Line

🗞 News of the day

🚀 AI BAN in the EU

The EU's AI Act is on the scene, and it's not playing around. By February 2025, companies must comply with new rules that classify AI systems into four risk levels: minimal, limited, high, and unacceptable. And if your AI falls into the "unacceptable" category, well, it's banned—totally, no ifs, ands, or buts.

So, what's considered "unacceptable"? Imagine AI systems doing creepy things like scoring you based on behavior, manipulating your decisions without you knowing, or targeting vulnerable groups—those are all no-gos. AI that uses biometrics to predict characteristics like sexual orientation? Not happening in the EU.

Companies that don't play by the rules could face fines up to a whopping €35 million or 7% of their annual revenue.

But don't stress just yet. The EU is giving everyone some breathing room until February 2025 to get their act together. Big names like Amazon, Google, and OpenAI have already jumped on the compliance bandwagon by signing the EU AI Pact, but others like Meta and Apple are still working on it.

There are some exceptions, though. Law enforcement can use biometric AI for targeted searches if they have the right permissions, and AI systems in workplaces or schools might get a pass if they're for medical or safety reasons. But don't get too excited—these exemptions come with strict conditions.

The industry is buzzing with concerns about how to navigate these new rules, especially with the overlap of other laws like GDPR. Companies are waiting on more guidance from the European Commission, and working groups are hustling to draft up codes of conduct.

🔖 Today's Cool Story

🚨 Major Breast Cancer Screening AI

AI breast cancer screening trial set to recruit 700,000 women in the UK.

• AI Platforms Tested: The NHS is launching a large-scale trial involving five AI platforms across 30 sites in the UK. This trial aims to improve the speed and efficiency of breast cancer diagnosis through AI technology, potentially freeing up radiologists for other tasks.

• Screening Process: Women aged 50 to 71, who are scheduled for routine NHS screenings, will be invited to participate in the £11 million Edith trial. Currently, two radiologists review mammogram images, but the new AI technology hopes to reduce this to one, allowing more patients to be seen and waiting lists to be shortened.

• Future of Cancer Care: Health Secretary Wes Streeting emphasizes the need for urgent action to improve cancer care in the UK. The government is seeking input from patients, staff, and experts to develop a national cancer plan. Dr. Katharine Halliday of the Royal College of Radiologists highlights that while AI has great potential, there is still a critical shortage of radiologists in the NHS.

💡 AI Insight

🎯 Uncrackable Optical Encryption AI

Researchers have developed an innovative encryption system combining holograms and AI to create nearly unbreakable optical security.

• Unbreakable Security: This new optical encryption system utilizes holograms to produce encryption that traditional methods cannot decipher. It enhances digital security for sensitive data by using a neural network that generates a unique decryption key, which can only be created by the system's owner.

• Advanced Technology: The system encodes laser beams with holograms, resulting in entirely scrambled light patterns that cannot be physically analyzed to reveal the original beam shape. Neural networks, trained to recognize these complex patterns, use billions of connections to reconstruct the original light shapes, with each system configuration having a specific decryption key.

• Future Prospects: Researchers plan to incorporate additional security measures such as two-factor authentication and are exploring cost-effective alternatives to high-power lasers. Despite challenges like cost and size, they aim to commercialize the technology for applications in cryptography and secure wireless optical communication.

🤖 Today on AI Robotics

🚨 OpenAI New Product Line

OpenAI is hinting at new product lines, including humanoid robots and AI-powered gadgets.

• OpenAI's Trademark Application: OpenAI has filed a trademark application with the U.S. Patent and Trademark Office, suggesting the development of user-programmable humanoid robots and other AI-powered hardware like headphones and smartwatches. The application also includes virtual and augmented reality headsets aimed at AI-assisted interaction. Some of these products might be launched soon, while others remain speculative.

• Collaboration with Jony Ive: OpenAI is collaborating with former Apple designer Jony Ive to create AI-powered gadgets. The collaboration aims to focus on innovative design and user experience. OpenAI's CEO Sam Altman mentioned that developing these prototypes may take several years, involving partnerships with multiple companies.

• Development of Humanoid Robots: OpenAI is working on humanoid robots designed for assistance and entertainment, led by Caitlin Kalinowski, formerly from Meta's AR glasses division. These robots will use custom sensors and AI for real-world operations, emphasizing intelligent interaction and learning capabilities. The potential applications for these robots include personal assistance and entertainment sectors.