The AI Jukebox Of Ads: Spotify’s India Play Could Make Radio Ads Obsolete

Here is what’s new in the AI world.

From Bollywood To Billions, Spotify’s Ad Exchange Takes India By Storm With AI Magic

Spotify has rolled out its Spotify Ads Exchange (SAX) and Generative AI-powered advertising tools in India, following successful pilots in the U.S. and Canada.

The move aims to streamline ad campaigns for brands while enhancing audience engagement on the platform.

Key Features:

Why It Matters?

SAX offers advertisers granular performance tracking and audience reach optimization while minimizing user ad fatigue. Over 5,000 global advertisers have tested the exchange, which Spotify’s India sales head Arjun Kolady says will simplify inclusion of Spotify in “always-on” programmatic strategies.

He emphasized Spotify’s goal to democratize ad creation for businesses of all sizes. The launch underscores Spotify’s push to monetize its growing Indian user base with tailored, tech-driven advertising solutions.

The Invisible Scam; AI’s Fake Job Seekers Are Stealing Your Opportunities

Organizations worldwide are facing mounting challenges in identifying fraudulent job applications, fueled by the rapid evolution of generative AI.

Sophisticated tools now enable scammers to craft fake headshots, clone voices, and fabricate credentials, while AI-generated resumes and cover letters further blur authenticity.

According to Gartner projections, deepfakes could account for 25% of global job applicants by 2028, signaling a looming crisis in hiring integrity.

Beyond Wasted Time: High-Stakes Risks

Fraudulent applicants pose severe threats, including malware installation, financial theft, and data breaches. Hiring teams must not only enhance scrutiny during recruitment but also prepare rapid response protocols for breaches.

The Detection Arms Race

While AI-powered tools like Pindrop’s deepfake detection platform aim to flag fake profiles, scammers are leveraging increasingly advanced AI to evade scrutiny.

Pindrop recently uncovered a Russian-linked scam operation using AI-generated identities to infiltrate companies, a stark irony, given the platform’s anti-deepfake design.

“Bad actors exploit synthetic identities, faces, and voices to fraudulently gain employment, forcing businesses to rethink verification frameworks”, he added.

As AI-driven fraud escalates, organizations must adopt multi-layered defenses, combining advanced detection tech with rigorous human oversight to safeguard against this invisible threat.

OpenAI Goes Open Source? Sam Altman Says Yes!

OpenAI CEO Sam Altman recently revealed plans to launch a powerful open-source AI model that surpasses anything currently available. While he acknowledged it could be used in ways some might not support, he emphasized the importance of open systems in the AI landscape.

Altman described the move as a timely step, enabled by a growing global understanding of AI risks, suggesting the company is now ready to contribute more openly.

This announcement came during a TED2025 interview with Chris Anderson, who asked whether the emergence of the Chinese open-source model DeepSeek had influenced OpenAI’s decision.

In response, Altman confirmed that OpenAI is preparing to release a highly capable open-source model, close to the performance of today's top-tier AI systems.

“It’ll be a very powerful system, likely the best open-source model available. We know not everyone will be happy with how it’s used, but open-source models have a rightful place in the broader AI ecosystem.”

Altman admitted that OpenAI had been slow to embrace open source but assured the audience that the company is now committed to doing it right.

Later in the interview, Altman revisited the topic with a touch of humor, suggesting that Anderson might criticize him next year if the open-source model is misused.

Still, he acknowledged the complexity of the decision and emphasized that OpenAI aims to release technology responsibly.

He elaborated: “We’ve been cautious because we weren’t sure how these systems would impact the world or how to keep them safe. That was a reasonable approach. But now that we better understand the landscape, it’s time to release more capable open systems.”

He continued, noting that while challenges and trade-offs will always exist, OpenAI’s goal remains to responsibly guide the development and deployment of AI technologies.

“We’ve mostly stayed true to our mission over the past decade. Our strategies will evolve, but our core mission remains strong.”

OpenAI’s move toward open source marks a significant shift. Although the upcoming model won’t be their most advanced, it’s expected to outperform all current open-source alternatives.

The company, which started as a nonprofit in 2015, became more closed-source in 2019 due to concerns about misuse. Altman’s comments reflect a renewed commitment to balancing innovation with responsibility.

While he didn’t say outright that OpenAI’s decision was prompted by DeepSeek’s popularity, his remarks suggest this move is a natural progression. He reiterated that open-source AI has a valuable and necessary role in the broader AI ecosystem.

Meet Reachy 2: The $70K Open-Source Robot You Can Hack

Hugging Face, a prominent player in the open-source AI space, has acquired Pollen Robotics, a French startup known for its open-source humanoid robots. This acquisition is aimed at pushing forward the accessibility and innovation of robotics in research, education, and real-world applications.

Led by former Tesla robotics engineer Remi Cadene, Hugging Face has been steadily expanding into robotics through its “Le Robot” initiative and its existing partnership with Pollen Robotics.

“We see robotics as the next major interface for AI and believe it should be open, affordable, and easy to modify,” the company stated.

Pollen Robotics, founded in 2016 by Matthieu Lapeyre and Pierre Rouanet, is best known for its humanoid robot Reachy.

The latest version, Reachy 2, is a modular, open-source robot equipped with advanced features like VR-based teleoperation, stereo vision, spatial audio, and a biologically inspired design.

Already adopted by universities like Cornell and Carnegie Mellon, Reachy 2 is available for purchase at $70,000 through the Pollen sales team.

Lapeyre commented on the company’s mission: “Our goal has always been to make robotics and AI approachable, engaging, and easy to use.”

He added that joining Hugging Face feels like a natural progression, as both companies share the same core values around accessibility and openness in technology.

With this acquisition, Hugging Face aims to further its commitment to open robotics, enabling developers and researchers to build custom robots without being locked into proprietary systems.

By combining its AI expertise with Pollen’s robotics hardware, Hugging Face plans to elevate Reachy’s functionality and support broader innovation in the field.

AI Takes Charge! Google Cloud's Bold Move for Enterprises

At Google Cloud Next 2025, the dominant theme was clear: AI is no longer a distant vision; it’s the driving force behind enterprise transformation today.

Beyond product unveilings, the event served as a blueprint for how businesses can embed AI into every aspect of their operations, from infrastructure to everyday productivity tools.

Google’s mission is to make AI broadly accessible, scalable, and seamlessly integrated with the tools companies already use. With this approach, Google is positioning itself not just as a provider of AI tools but as a key partner in enterprise digital transformation.

“AI is the biggest opportunity we’ve seen,” said Sundar Pichai, CEO of Google and Alphabet. “That’s why we’re investing across the entire AI technology stack, starting with the foundational infrastructure.”

Google plans to allocate approximately 75 billion USD in capital expenditure in 2025 to support AI computing and cloud services.

Gemini and Vertex AI: Enterprise AI Made Practical

At the heart of Google’s AI ecosystem is the Gemini model family. Gemini 2.5 Pro supports complex reasoning and decision-making, while Gemini 2.5 Flash offers speed and cost-efficiency, ideal for customer-facing and real-time applications.

These models are deployed through Vertex AI, Google’s enterprise AI platform, which now includes access to over 200 models, including Google’s own tools like Imagen (text-to-image), VEO (long-form video), and Lyria (music generation), along with popular models from Meta and AI21 Labs.

Vertex AI has received major updates, new dashboards for real-time performance tracking, enhanced model tuning for optimization, and improved model grounding, which ties outputs to trusted sources like Google Search, Maps, and enterprise databases.

It also allows seamless integration with platforms like SAP, Oracle, and ServiceNow, eliminating the need to duplicate or restructure existing data.

AI Agents: The Next Evolution in Digital Workflows

A major highlight was the debut of Google’s AI agent ecosystem, marking a major step forward in AI’s role in the workplace. With the Agent Development Kit (ADK), an open-source framework, businesses can now build intelligent agents that plan, reason, and even collaborate.

These agents live inside Google Agentspace, a platform designed for enterprise use, where employees can use them to automate tasks, retrieve information, and improve operational efficiency. They integrate with both internal and external data, creating a flexible AI-powered layer across business operations.

This represents a shift from AI being a specialized tool for data teams to a productivity booster accessible to every employee.

Next-Gen Infrastructure for AI Workloads

To support the increasing demand for AI, Google introduced major upgrades to its compute infrastructure. The seventh-gen Tensor Processing Unit (TPU), Ironwood, delivers up to 3,600 times the performance of the original TPU and can provide 45 exaflops of compute per pod.

In addition, Google Cloud now supports NVIDIA’s most advanced GPUs, including the Grace Blackwell GB200 and the upcoming Rubin series.

All of this is part of the AI Hypercomputer; Google’s next-gen supercomputing platform combining elite hardware and integrated software to improve speed, performance, and cost-efficiency.

Adding to this, Google’s Cloud WAN offers a global, private network delivering up to 40% performance improvements and significant cost savings—essential for running distributed, low-latency AI workloads globally.

AI Embedded In Daily Work: Google Workspace

Google is also embedding AI directly into its productivity suite. Gemini is now fully integrated into Google Workspace, enhancing Gmail, Docs, Sheets, and Meet with intelligent features.

For example:

These features put powerful AI tools directly into the hands of employees, driving better decisions and boosting productivity at all levels.

Real-World Success Across Industries

Google shared examples of leading companies already benefiting from these AI tools:

• McDonald’s uses Gemini to improve restaurant operations in real-time.

• Intuit is transforming tax prep with Document AI.

• Deutsche Bank leverages AI agents for financial research.

• Verizon uses AI assistants to enhance customer service.

These case studies demonstrate how AI is already driving meaningful results, cutting costs, improving speed, and enabling new services.

What It Means for Business and ERP Leaders

Google Cloud Next 2025 made one thing clear: AI is now essential, not optional. Tools like Gemini, Vertex AI, and Workspace integrations show that enterprises can implement AI across operations today to enhance productivity, decision-making, and customer engagement.

Google has also made AI infrastructure, like advanced TPUs, GPUs, and the Hypercomputer, more accessible, ensuring businesses of all sizes can scale AI effectively and securely.

And with intelligent agents now on the scene, a new workplace model is emerging, one where AI helps automate, support, and amplify the efforts of every employee.

Your opinion matters!

Hope you loved reading our piece of newsletter as much as we had fun writing it. 

Share your experience and feedback with us below ‘cause we take your critique very seriously. 

Thank you for reading

-Shen & Towards AGI team