OpenAI Issues Ban Warnings to Users Investigating ‘Strawberry’ AI Models

A Thought Leadership platform to help the world navigate towards Artificial General Intelligence We are committed to navigate the path towards Artificial General Intelligence (AGI) by building a community of innovators, thinkers, and AI enthusiasts.

Whether you're passionate about machine learning, neural networks, or the ethics surrounding GenAI, our platform offers cutting-edge insights, resources, and collaborations on everything AI.

What to expect from Towards AGI: Know Your Inference (KYI): Ensuring your AI-generated insights are accurate, ethical, and fit for purpose Open vs Closed AI: Get expert analysis to help you navigate the open-source vs closed-source AI GenAI Maturity Assessment: Evaluate and audit your AI capabilities Expert Insights & Articles: Stay informed with deep dives into the latest AI advancements But that’s not all!

We are training specialised AI Analyst Agents for CxOs to interact and seek insights and answers for their most pressing questions. No more waiting for a Gartner Analyst appointment for weeks, You’ll be just a prompt away from getting the insights you need to make critical business decisions. Watch out this space!!

Visit us at https://www.towardsagi.ai to be part of the future of AI. Let’s build the next wave of AI innovations together!

TheGen.AI News

OpenAI Issues Ban Warnings to Users Investigating ‘Strawberry’ AI Models

OpenAI is keeping the inner workings of its new "Strawberry" AI model family under wraps. Since launching models like o1-preview and o1-mini last week, which boast enhanced reasoning abilities, OpenAI has been issuing warnings and threatening bans to users attempting to uncover how the model operates.

Unlike earlier models such as GPT-4o, o1 is designed to go through a step-by-step problem-solving process before generating an answer. While users can view this process in the ChatGPT interface, OpenAI intentionally hides the raw reasoning process, instead displaying a filtered version created by a second AI model.

This secrecy has sparked curiosity among enthusiasts, leading hackers and red-teamers to try jailbreaking or prompt injection techniques to reveal the raw reasoning. Some have reportedly had success, but no solid confirmation has emerged. OpenAI, monitoring through the ChatGPT interface, has responded aggressively to such efforts.

Several users, including Scale AI’s prompt engineer Riley Goodside, reported receiving warning emails for simply mentioning terms like "reasoning trace" when using o1. The warnings indicate that the flagged activity violates OpenAI's policies on bypassing safeguards and urge users to stop. The email warns that further violations could result in the loss of access to the GPT-4o model with reasoning capabilities.

Marco Figueroa, who manages Mozilla’s GenAI bug bounty programs, was among the first to share details about OpenAI’s warning on X. He expressed frustration that it interferes with his ability to conduct safety research through red-teaming, noting that after several jailbreak attempts, he ended up on OpenAI’s “get banned” list.

LinkedIn Activates GenAI Training with Personal Data by Default

LinkedIn has quietly introduced a new feature that uses users' personal data to train Generative AI (GenAI) models, which is activated by default. This affects a large portion of LinkedIn's 800 million users, as LinkedIn, Microsoft, and its affiliates can now use their data to train AI systems. Notably, this data collection excludes users in the EU, EEA, and Switzerland, likely due to the EU AI Act's data governance rules.

LinkedIn hasn't clarified whether direct messages are included in the data used, but they state that they aim to minimize personal data in the training process using privacy-enhancing technologies. The platform’s updated terms indicate that LinkedIn collects data related to users’ activity, such as posts, language preferences, and feedback, and uses it for generative AI features.

In addition to LinkedIn’s own models, external providers like Microsoft’s Azure OpenAI service may also utilize the data. This exclusion of European users could be due to the stricter privacy laws in the region, which require opt-in consent for data use.

Heather Burns, a tech policy expert, suggested that LinkedIn’s decision to exclude Europe may reflect concerns about its GenAI model’s accuracy in the face of strict privacy regulations. In the UK, where LinkedIn is still required to obtain opt-in consent, scrutiny from the Information Commissioner's Office (ICO) is likely.

Legal experts warn that LinkedIn’s opt-out model could lead to backlash from both users and regulators. There are concerns that allowing LinkedIn to use data for AI training could expose businesses to risks, such as leaking proprietary information and business strategies. This could harm a company's reputation if clients perceive it as neglecting data protection.

Experts also emphasize the importance of emerging encryption technologies, like Fully Homomorphic Encryption (FHE), that could help businesses use AI responsibly while complying with global data protection laws.

Amazon Unveils Project Amelia: A GenAI Assistant for Sellers

At Amazon's Accelerate event in Seattle on September 19, the company unveiled a new Generative AI (GenAI) initiative designed to simplify the complexities faced by independent sellers using its platform. Known as Project Amelia, this personal assistant aims to help sellers streamline their operations and scale their businesses more efficiently.

Mary Beth Westmoreland, Amazon’s VP of Worldwide Selling Partner Experience, emphasized a practical approach to technology, stating that it should deliver real-world applications rather than being used just for its own sake. With millions of third-party sellers on the platform, Amazon's marketplace represents over 60% of its eCommerce sales, and the Accelerate event introduced a variety of tools and upgrades to support sellers.

Westmoreland explained that operating a global selling business on Amazon involves many tasks, from product development to inventory management, and while Amazon already provides tools to assist sellers, Project Amelia is designed to make these processes even more straightforward. During a demo, Project Amelia showed its capability to respond to general seller queries by providing intelligence reports with tailored advice on inventory or logistics.

For example, if an apparel seller asks about expanding their product lines, Amelia offers recommendations based on the seller’s existing inventory and market trends. It can also generate detailed status reports, covering revenue, year-to-date performance, and comparisons with projections. The AI assistant is accessible across all pages in Amazon’s Seller Central platform, where sellers manage their businesses.

Built on Amazon Bedrock, which supports scalable GenAI applications, Project Amelia combines general world knowledge with Amazon-specific expertise to provide unique, actionable insights. Over time, the AI will evolve to offer more personalized guidance, anticipating sellers’ needs and resolving issues proactively.

Project Amelia focuses on three main areas: answering knowledge-based questions, updating business metrics, and resolving issues. Sellers can ask for advice on topics like holiday preparation strategies or compliance, check performance metrics like sales and customer traffic, and, in the future, even resolve inventory discrepancies through automated actions.

Currently in beta, Project Amelia is available to select U.S. sellers, with plans for broader U.S. access and support for more languages by the end of the year. As the AI assistant develops, Amazon envisions Amelia playing a more proactive role in managing sellers' operations, ultimately freeing up time for business growth. Westmoreland sees this as the beginning of Amelia’s evolution into a comprehensive, trusted advisor for Amazon sellers.

Nasscom Nurtures 37 Indian GenAI Startups in its Latest Innovation Cohort

Nasscom, the leading industry body for IT in India, announced on Thursday the selection of 37 Generative AI (GenAI) startups for the second cohort of its ‘Generative AI Foundry’ programme, with 30% of these startups having women in their founding teams. Notable companies in the cohort include Fluid AI, Sivi, DaveAI, and OnFinance. The selected startups, on average two years old, have collectively raised $750,000 from institutional investors.

“Indian GenAI startups are not only transforming industries domestically but are also making a global impact, highlighting India’s growing AI expertise on the world stage,” said Ankit Bose, Head of AI at Nasscom.

Chosen from over 130 applicants, these startups are innovating in areas such as text, audio, code, video/image generation, and multimodal outputs. They are developing solutions with strong applications in sectors like retail, banking and financial services, healthcare, life sciences, and enterprise operations.

The ‘Generative AI Foundry’ programme, launched in October 2023, currently supports 26 startups and has partnerships with nearly 60 enterprises and investors. The programme has seen business growth double and staffing increase by 50%, collectively raising $15 million in funding, with an average of $2 million expected per startup.

Nasscom noted that these startups have filed over 40 patents, with more in progress, highlighting significant intellectual property creation in India. Currently, more than 200 tech startups in India are actively developing and offering GenAI solutions across various industries.

According to a recent Nasscom-Zinnov report, deeptech funding in 2023 exceeded $500 million. However, the industry body emphasized the importance of increasing collaboration between startups and industry, improving access to affordable GPUs, patient capital, applied research, and building a strong talent pipeline to foster the growth of the GenAI startup ecosystem in India.

TheOpensource.AI News

Alibaba Unveils 100+ Open-Source AI Models and Text-to-Video Tool

Alibaba announced the release of over 100 open-source AI models, alongside enhancements to its proprietary technology, as it seeks to bolster its competitive edge against rivals. The new models, dubbed Qwen 2.5, are designed for various sectors such as automotive, gaming, and scientific research, with advanced math and coding capabilities.

Based in Hangzhou, Alibaba is aiming to compete more aggressively with domestic players like Baidu and Huawei, as well as global giants like Microsoft and OpenAI. AI models rely on vast amounts of data for training, and Alibaba claims its models can understand prompts and generate text and images. 

By open-sourcing the Qwen models, Alibaba allows researchers, academics, and companies worldwide to use them for building their own generative AI applications without needing to train their own systems, which saves both time and costs. Alibaba hopes this approach will attract more users to its AI technology.

The company first launched its Tongyi Qianwen, or Qwen, model last year and has since released improved versions. To date, its open-source models have been downloaded 40 million times. Alibaba also upgraded its proprietary Qwen-Max model, which is not open-source and is sold through the company’s cloud computing services. Alibaba claims that Qwen Max 2.5-Max outperforms competitors like Meta's Llama and OpenAI's GPT-4 in areas such as reasoning and language comprehension.

Additionally, Alibaba introduced a new text-to-video tool based on its AI models, which allows users to create videos from prompts, similar to OpenAI’s Sora. 

Eddie Wu, CEO of Alibaba, emphasized the company’s significant investment in AI technology and global infrastructure development. Wu, who became CEO last year during a major leadership reshuffle, is working to reignite growth at the company amid increasing competition and a sluggish Chinese market. Although Alibaba is a dominant cloud computing player in China, it trails behind Amazon and Microsoft internationally. The company hopes its latest AI offerings will attract more customers, both in China and abroad, to its cloud services, which have shown early signs of acceleration in the recent June quarter.

Open Source AI Crucial for Europe’s Digital Sovereignty and Growth

AI and open source are vital to the future of Europe’s technology sector, particularly for the region’s digital sovereignty and growth opportunities through open source investments. This was a key focus during the Open Source Summit Europe, where the Linux Foundation’s “Open Source Maturity in Europe” report was released.

Gabriele Columbro, General Manager of the Linux Foundation Europe and Executive Director of the Fintech Open Source Foundation, emphasized the urgency, stating, “Europe cannot afford to miss another technological wave. This goes beyond economics—it’s about digital sovereignty.”

According to the report, 43% of respondents believe that artificial intelligence and machine learning (AI/ML) would gain the most from being open source, underscoring the importance of transparent AI development in Europe. AI/ML was ranked higher than other technologies like operating systems (40%), cybersecurity (30%), and cloud or container technologies (29%).

The survey also highlighted the growing priority of investing in open source as an alternative to tech monopolies, with 58% of respondents identifying it as the top priority for 2024, up from 39% in 2023. Additionally, security perceptions have shifted, with 76% of respondents considering open source software more secure than proprietary alternatives.

Despite growing confidence in open source, Europe still lags behind the U.S., particularly in the cloud sector, where it lacks competitors to giants like Google Cloud, AWS, and Azure. While companies like Mistral AI and Spotify are driving support for open source, Europe faces challenges in building alternatives, especially in the cloud.

Columbro pointed out that discussions around digital sovereignty are intensifying, particularly regarding the cloud, but efforts to establish a European cloud have often failed. He noted that once dominant platforms like Kubernetes gain momentum, it becomes much harder to create alternatives, stressing that the timing to act is critical.

TheClosedsource.AI News

OpenAI Taps Former Coursera Exec to Lead AI Integration in Education

OpenAI has appointed Leah Belsky, former executive at Coursera Inc., as its first general manager of education to expand its AI tools into more schools and classrooms. Belsky, who was the chief revenue officer at the online learning platform, now steps into the role of engaging instructors and students across K-12, higher education, and continuing education. Her mission: drive OpenAI’s outreach efforts and strengthen partnerships with academic institutions while relaying critical feedback from educators.

Since the release of ChatGPT in late 2022, students quickly embraced the tool, resulting in a noticeable surge in usage at the start of the school year. Some teachers initially pushed back, concerned about AI’s potential to foster cheating. But many others began integrating the technology into their lesson plans and research. OpenAI responded by developing deeper connections with academia. In May, the company introduced ChatGPT Edu, a specialized version with enhanced controls and tailored pricing for educational institutions. Universities such as Oxford, Arizona State, and Columbia have already started using the platform.

Brad Lightcap, OpenAI’s chief operating officer, emphasized Belsky’s role in accelerating collaborations with leading academic institutions. “Leah will drive our partnerships to ensure people in all fields have the AI training necessary to harness its full potential,” he said.

OpenAI is set to host an October meeting with university presidents and provosts, aiming to discuss responsible AI adoption for teaching and research. As the company's AI models grow more advanced, their capabilities in fields like physics, chemistry, and biology have begun to rival those of Ph.D. students on certain academic benchmarks.

Unlock the future of problem solving with Generative AI!

If you're a professional looking to elevate your strategic insights, enhance decision-making, and redefine problem-solving with cutting-edge technologies, the Consulting in the age of Gen AI course is your gateway. Perfect for those ready to integrate Generative AI into your work and stay ahead of the curve.

In a world where AI is rapidly transforming industries, businesses need professionals and consultants who can navigate this evolving landscape. This learning experience arms you with the essential skills to leverage Generative AI for improving problem-solving, decision-making, or advising clients.

Join us and gain firsthand experience in how state-of-the-art technology can elevate your problem solving skills using GenAI to new heights. This isn’t just learning; it’s your competitive edge in an AI-driven world.

🦾 Master AI & ChatGPT for FREE in just 3 hours 🤯

1 Million+ people have attended, and are RAVING about this AI Workshop.
Don’t believe us? Attend it for free and see it for yourself.

Save your spot here. (100 free spots only)

Highly Recommended: 🚀

Join this 3-hour Power-Packed Masterclass worth $399 for absolutely free and learn 20+ AI tools to become 10x better & faster at what you do

👉Save your seat  now (FREE for First 100)

🗓️ Tomorrow | ⏱️ 10 AM EST

In this Masterclass, you’ll learn how to:

🚀 Do quick excel analysis & make AI-powered PPTs 
🚀 Build your own personal AI assistant to save 10+ hours
🚀 Become an expert at prompting & learn 20+ AI tools
🚀 Research faster & make your life a lot simpler & more…

👉 Register here (Offer valid for First 100 people only)🎁

In our quest to explore the dynamic and rapidly evolving field of Artificial Intelligence, this newsletter is your go-to source for the latest developments, breakthroughs, and discussions on Generative AI. Each edition brings you the most compelling news and insights from the forefront of Generative AI (GenAI), featuring cutting-edge research, transformative technologies, and the pioneering work of industry leaders.

Highlights from GenAI, OpenAI, and ClosedAI: Dive into the latest projects and innovations from the leading organizations behind some of the most advanced AI models in open-source, closed-sourced AI.

Stay Informed and Engaged: Whether you're a researcher, developer, entrepreneur, or enthusiast, "Towards AGI" aims to keep you informed and inspired. From technical deep-dives to ethical debates, our newsletter addresses the multifaceted aspects of AI development and its implications on society and industry.

Join us on this exciting journey as we navigate the complex landscape of artificial intelligence, moving steadily towards the realization of AGI. Stay tuned for exclusive interviews, expert opinions, and much more!

Keep reading