This Week In AI (8th Feb 2025)
Ethical Dilemmas, Model Breakthroughs, and the Future of Smart Glasses
This week, AI took center stage with major launches, ethical controversies, and security innovations. Let’s break it all down.
PS: The first 3 editions of “This Week in AI” newsletter will be FREE 🥳
OpenAI Unveils GPT-4’s Successor: O3 Mini and DeepResearch
Here’s What You Need to Know:
OpenAI has released O3 Mini, a smaller, optimized version of its latest model, alongside DeepResearch, a specialized AI built for in-depth academic and scientific work. O3 Mini promises efficiency and accessibility, while DeepResearch aims to elevate AI’s role in research and knowledge generation.
Why It’s Important for AI Professionals:
These releases mark OpenAI’s shift towards lightweight, domain-specific models, signaling a future where AI becomes more modular and purpose-driven. Developers and researchers should expect more task-optimized models that integrate seamlessly into workflows.
Why It Matters for Everyone Else:
DeepResearch could revolutionize academic fields, making high-quality AI-assisted research accessible to scientists and students. Meanwhile, O3 Mini’s efficiency means more powerful AI in consumer applications with lower computing costs.
Aish’s Prediction:
I got ChatGPT Pro for Operator, but I am staying on it for DeepResearch. I absolutely love using DeepResearch, as it is able to do a ton of market research for me. I am able to build reports in any format that I need including documents, tabular summary, or even mind-maps! This is totally worth the 200$ subscription (but, I feel the Pro subscription will become cheaper in the future!). My prediction is that a ton of people are going to use DeepResearch more than Operator for learning purposes.
AI Copyright Laws: Who Owns AI-Generated Content?
Here’s What You Need to Know:
Global discussions on AI-generated content are intensifying as governments and courts debate copyright laws. Some propose that AI-created works belong to the model’s developer, while others argue they should remain in the public domain.
Why It’s Important for AI Professionals:
Copyright laws will directly impact dataset access, model training, and the legal standing of AI-generated outputs. Understanding these regulations will be crucial for researchers and businesses deploying generative AI.
Why It Matters for Everyone Else:
This ruling could reshape media, journalism, and creative industries. If AI-generated content is proprietary, companies may face licensing fees or restrictions, while independent creators may struggle to protect their work.
Aish’s Prediction:
AI copyright laws are about to get messy. Right now, we’re in a legal gray zone where no one really knows who owns AI-generated content—governments are scrambling, courts are debating, and tech companies are (quietly) lobbying for outcomes that work in their favor.
Here’s my take: We’re going to see a patchwork of regulations emerge globally. Some countries will push for AI-generated works to be public domain (which would be a win for open innovation), while others will favor giving ownership to developers or the companies behind the AI models.
Biggest impact? Businesses training AI models will face more legal hoops to jump through, and creators will have to navigate licensing nightmares. If AI-generated content becomes proprietary, expect lawsuits, paywalls, and some very interesting copyright battles ahead.
My bet? By the end of this year, at least one major AI model will face a high-profile legal challenge over content ownership. Popcorn ready. 🍿
Google Quietly Removes Ethical AI Clause on Surveillance
Here’s What You Need to Know:
Google has removed its internal policy prohibiting AI development for surveillance applications. This change has sparked concern among ethicists and privacy advocates, given AI’s potential use in mass monitoring and government oversight.
Why It’s Important for AI Professionals:
This signals a shift in corporate AI ethics, with potential ripple effects on industry standards. AI engineers working in large companies should prepare for increasing pressure to balance innovation with responsible development.
Why It Matters for Everyone Else:
With fewer safeguards, AI-powered surveillance could expand rapidly, affecting privacy rights, workplace monitoring, and law enforcement practices. The balance between security and civil liberties will become a key societal debate.
Aish’s Prediction:
Google dropping its ethical AI clause on surveillance? Yeah… that’s not a good sign. This isn’t just about one company—it’s a signal that AI ethics in Big Tech are becoming more flexible (read: optional).
Here’s why I see this is happening: The original Responsible AI principles that companies put together were future-looking and they didn’t anticipate Generative AI to be adopted so widely so fast. Now that the companies are realizing “business -value” in what they thought would be perceived unethical, they are revising their principles. Google is not the only company, I am pretty sure several companies are doing the same- just not making the headlines!
The Future of Smart Glasses: Are They Finally Ready for Prime Time?
Here’s What You Need to Know:
Smart glasses have seen waves of hype but struggled with adoption. Now, with advancements in AI-powered AR overlays and improved battery life, companies like Meta and Apple are betting big on their mainstream appeal.
Why It’s Important for AI Professionals:
AI-powered smart glasses will require robust vision models, real-time processing, and efficient UX design. Engineers focusing on AR/VR should watch this space as it shifts from novelty to productivity tool.
Why It Matters for Everyone Else:
Beyond fashion accessories, smart glasses could redefine how we interact with digital content, from seamless translations to real-world AR shopping and productivity enhancements.
Aish’s Prediction:
I’ve been using my Meta Ray-Ban glasses, and I love them. The experience is seamless, and when smart glasses look normal and feel natural, they just work.
If companies can nail style, affordability, and AI-powered utility, this won’t just be a niche product—it’ll be a hit. My bet? Once pricing makes sense for everyday users, we’ll see a wave of adoption. Smart glasses aren’t just the future—they’re finally about to have their moment.
Anthropic Develops New Defense Against AI Jailbreaks
Here’s What You Need to Know:
Anthropic has introduced a novel method to prevent large language models from being jailbroken, addressing a key security concern in AI safety. This innovation could reduce misuse cases where models are tricked into bypassing restrictions.
Why It’s Important for AI Professionals:
Stronger jailbreak defenses mean AI developers must rethink prompt engineering techniques and adversarial testing. Model alignment will remain a key challenge in security-sensitive applications.
Why It Matters for Everyone Else:
As AI becomes more integrated into consumer applications, these defenses are crucial for ensuring AI systems don’t provide harmful or unethical responses.
Aish’s Prediction:
AI jailbreaks are a constant cat-and-mouse game, and Anthropic just made a strong move on defense. Their new method will make it harder to trick models into bypassing restrictions—but let’s be real, hackers and prompt engineers will immediately start finding new exploits.
Stronger guardrails are great for safety, but they also make AI models feel more restricted, which can frustrate users. My bet? We’ll see an ongoing tug-of-war: companies tightening security while power users push for more flexibility. And in the next year, at least one major AI system will face a high-profile jailbreak that forces companies to rethink their approach again.
Final Thoughts
AI continues to push boundaries—sometimes for the better, sometimes raising ethical questions. As we navigate this evolving landscape, staying informed is our best tool for shaping AI’s future responsibly.
What do you think about this week’s AI developments? Let’s discuss!
PS: This newsletter is AI-generated every week that fetches the more recent news based on a detailed set of prompts that Aish provides. Once the newsletter is ready, Aish adds her thoughts and predications personally to each of the news pieces.