Apple AI‘s Huge Bug

Hello, AI Enthusiasts!

In today's issue, we dive into a gripping story that highlights the delicate balance between cutting-edge technology and security.

AI Secret helps you grasp the latest AI trends and insights in just 5 minutes. Don’t remember subscribing or want to stop receiving our newsletters completely? Unsubscribe here

TODAY’S AI SECRET:

• A developer manipulated Apple Intelligence.
• Intel is going through a series of major failures.
• Daily trending & featured AIs boost your career and business.
• Publishers make millions, while authors get nothing.

• OpenAI Collaborates with The Met - The partnership introduces an interactive chat experience with a representation of socialite Natalie Potter, enhancing visitors' engagement with the "Sleeping Beauties" exhibit.
• AI Experiment Raises Alarms - ChaosGPT, an open-source AI, attempts to "destroy humanity" through tweets and research in a controversial demonstration.
• Microsoft Crashed Again - GitHub announces that its services are now “fully operational” after a major outage caused by database changes. Since Microsoft's acquisition in 2018, some users have doubted its service stability.
• Cisco Layoffs Amid Shift to AI - Cisco announces a 7% reduction in workforce, laying off approximately 5,900 employees as it pivots towards AI and cybersecurity.

A developer manipulated Apple Intelligence.

Recently, just as Apple was poised to release its latest AI functionality—Apple Intelligence—a developer named Evan Zhou revealed a significant security flaw in the system. Zhou then published his code on GitHub.

The Details:

• Security Flaw Discovery: Evan Zhou demonstrated in the beta version of MacOS 15.1 how simple code could manipulate Apple Intelligence.
• Method of Manipulation: Zhou showcased a method where malicious commands are injected to manipulate AI behavior. This successful attack highlights the vulnerability in how AI systems handle input and instructions, due to the blurred boundaries between them.
• Technical Background and Impact: Zhou's case is not isolated but reflects a broader issue. OWASP has also listed prompt injection as one of the primary threats to large language models. When attackers can manipulate AI with simple inputs, causing it to leak sensitive information or perform malicious actions, the trust and security of the entire system are severely compromised.

Why It Matters:

Evan Zhou's example serves as a reminder that, alongside technological innovation, security cannot be overlooked. For companies and developers, ensuring that AI systems are resilient against such attacks is crucial for protecting user privacy and data security.

The #1 Daily Audio Book Summary APP

"Busy life, but love to read? Transform your every moment into a learning opportunity with HeardLy APP!

Whether you're commuting, working out, or just relaxing, HeardLy lets you dive into the world of books effortlessly.

With our AI-driven app, quickly master a top book in just a few minutes.

Sign up for a free trial today and gain 30-day access to master 30 top books!

Don't just fill your time—enrich it with HeardLy."

Intel is going through a series of major failures.

SoftBank and Intel's plan to create an AI chip to rival Nvidia has been scrapped. Intel couldn't meet SoftBank's volume and speed needs, prompting SoftBank to shift its focus to TSMC.

The Details:

• Partnership Breakdown: Intel failed to meet SoftBank's specific demands for the AI chip production, causing the partnership talks to collapse.
• Impact on Intel: The failed negotiations preceded Intel's significant cost-cutting measures, including widespread layoffs.
• Shift in Strategy: SoftBank is now engaging with TSMC, aiming to leverage the capabilities of the world's largest contract chipmaker for its AI chip ambitions.

Why It Matters:

Right now, Intel, once a leading chip maker, appears to be struggling. After facing several challenges, such as problems with new chips, a drop in stock prices, and significant layoffs, even new companies like SoftBank have decided to distance themselves from Intel. I can't help but wonder if Intel will end up like Nokia?

• 🔑 Website Summarizer captures key insights of any website in 5 seconds.
• 📋 CareerCompanion generates interview questions from a job description.
• 📝 Sheet Copilot uses AI to complete tasks in Google Sheets.
• 📊 Hepta uses AI to summarize your data quickly and efficiently.
• 🪪 Onfido creates trust at onboarding and beyond with an AI-powered digital identity solution.
• 🚀 Flot AI: Write, Read, Memorize, with Al at your side. (Featured)
• 💡 MyMap AI transforms Ideas into Visual Stories. (Featured)
• 🥷 BypassGPT generates AI content undetectable by AI detection. (Featured)

Publishers make millions, while authors get nothing.

A recent disturbing practice in the academic world: publishers selling academic papers to tech companies for AI training, earning millions while the original authors receive nothing. This raises significant ethical and legal questions about the use of intellectual property and the compensation of creators.

The Details:

• Exploitative Earnings: Publishers like Taylor & Francis and Wiley have made substantial profits, up to $23 million, by licensing academic papers to tech companies for AI training.
• Authors' Lack of Compensation: Despite their hard work, authors of these papers have not received any financial benefits from the commercial use of their work.
• Legal and Ethical Concerns: The practice has sparked debates about the legality and ethics of using academic research in this way, especially without compensating the authors.

Why It Matters:

As AI continues to evolve, the need for massive datasets has led to the commodification of academic research, often at the expense of the original creators. This practice not only questions the fairness of the academic publishing model but also the broader implications for how we value and protect intellectual contributions in the digital age.

1 Million+ AI enthusiasts are eager to learn about your product.

AI Secret is the world’s #1 AI Newsletter, boasting over 1 million readers from leading companies such as OpenAI, Google, Meta, and Microsoft. We've assisted in promoting Over 500 AI-Related Products. Will yours be the next?

What We Can Offer:

• Launch an Advertising Campaign
• Introduce New Product or Features
• Other Business Cooperation

Email our co-founder Mark directly at mark@aisecret.us if the button fails.