AI Bytes Newsletter Issue #12

For more catch us at antics.tv!

Welcome to the 12th edition of the AI Bytes newsletter, where we spotlight the collaboration between ElevenLabs and rabbit to upgrade the r1 device with lifelike voice capabilities, emphasizing our commitment to making technology more intuitive and human-centered. This edition also examines a cautionary tale where a company was tricked out of $25 million through a sophisticated deepfake scam, highlighting the urgent need for heightened security and ethical awareness in AI applications. We introduce BRANDMARK as our AI Tool of the Week, offering a glimpse into the future of effortless logo creation powered by AI. Rico's Roundup navigates the complex landscape of AI in public safety and governance, advocating for a balanced approach to technological integration. Mike's Musings simplifies the intricate concepts of deterministic and non-deterministic AI, providing clear examples to illustrate these foundational AI principles. This edition is packed with insights aimed at enriching your understanding of AI's evolving role in our world.

The Latest in AI

A Look into the Heart of AI

Featured Innovation

ElevenLabs gives lifelike voice to rabbit’s r1 device

The collaboration between ElevenLabs and rabbit is bringing about an upgrade to the r1 device, focusing on making interactions more intuitive through advanced voice AI. This effort aims to bridge the gap between human and machine communication, making the technology more relatable and easier to use.

By incorporating ElevenLabs' voice technology, the goal is to enable the r1 device to not just receive commands but engage in more natural conversations. This marks a step forward in making our interactions with devices less mechanical and more like talking to another person.

This initiative is part of the wider effort to integrate AI more seamlessly into everyday life. It reflects the ongoing trend towards creating technology that's not only functional but also more accessible and engaging.

Discussions about such technological advancements are essential. They help us understand the potential of AI to improve our daily interactions with technology, while also considering the ethical implications of increasingly human-like AI systems.

Check out the link below for more information and the official 11 blog post:

If you’ve seen a game changing innovation and want to share it with us, hit us up at [email protected]!

Ethical Considerations & Real-World Impact

In a scenario that seems plucked straight from the pages of a science fiction novel, a multinational company was swindled out of a staggering $25 million through an elaborate deepfake scam. This high-tech heist involved the use of sophisticated deepfake technology to mimic the company's Chief Financial Officer (CFO) during a video conference call, a tactic so convincing it would make any Hollywood special effects team proud.

The scam began when a finance worker, unsuspecting and diligent, was lured into what was believed to be a routine video call with colleagues. However, the reality was far more sinister. The "colleagues" present in the meeting were nothing more than digital fabrications, deepfake recreations engineered to deceive and manipulate. This wasn't just a breach of digital security; it was a stark violation of trust and a glimpse into a dystopian potential future where seeing might no longer be believing.

The unsuspecting employee, convinced by the digital masquerade, was manipulated into transferring 200 million Hong Kong dollars (approximately $25.6 million), under the guise of executing a secretive transaction at the behest of the impersonated CFO. This incident wasn't just a financial loss; it was a chilling demonstration of the power and potential misuse of deepfake technology.

The sophistication of this scam underscores a critical vulnerability in our increasingly digital world: the delicate balance between technological advancement and ethical application. As deepfake technology becomes more accessible and indistinguishable from reality, the potential for misuse in social engineering scams becomes alarmingly clear. This incident serves as a wake-up call, highlighting the urgent need for advanced security measures, awareness, and training to combat such high-tech deceptions.

Tools

The Toolbox for Navigating the AI Landscape

AI Tool of the Week - BRANDMARK (brandmark.io)

It isn’t every day that you need to make a logo (or maybe you’re a graphic designer and it is!), and when you do need a logo, the process of design iteration etc. can be costly and time consuming. Even going on sites like Fiverr will likely have you going through multiple design iterations which can take time when you’re waiting for a human on the other end.

Enter brandmark.io. Brandmark is a simple logo generating service, it’s free to get started, in order to use the logo, there’s a small cost.

Here’s a direct link to brandmark’s logo generation tool: https://app.brandmark.io/v3/. If you try this out, let us know what you think!

If you’ve got a suggestion on tools we should check out, email us at [email protected] and let us know.

Rico's Roundup

Critical Insights and Curated Content from Rico

Skeptic's Corner: Balancing Act — Negotiating the Promise and Perils of AI in Public Safety and Governance

In this week's edition of the Skeptic's Corner, I am doing things a bit different. Below, in the "Must-Read Articles" section, you'll find two pieces focusing on the implementation of AI across two distinct yet equally significant spheres: the New York City Subway system and the broader adoption of AI within US Federal Agencies. As the designated house skeptic, it would be negligent of me not to voice my persistent apprehensions regarding AI systems' role in policing the populace. It's a sentiment I'm sure many of our readers and listeners can relate to, having been fed a steady diet of cautionary tales in movies where AI or technology veers off course, often with disastrous consequences for humanity or the planet at large (or freedom).

While I acknowledge and appreciate the value of employing AI to streamline federal operations—thinking about chatbots aiding in customer service, algorithms sifting through vast datasets to extract actionable insights, or systems that minimize human error and amplify our capabilities—I have serious reservations about the widespread use of facial recognition and firearm detectors. These technologies, despite their potential for enhancing safety and efficiency, carry a significant risk of misidentification and misuse. This concern is exacerbated by the current legislative void and a perceptible disconnect between government actions and public sentiment. Safety, while of utmost importance, should not be pursued without a thorough and fully-transparent discussion on how such systems are deployed among the populace. The delicate balance between modern conveniences, safety enhancements, and the preservation of freedoms as originally intended, stands as a central theme in our discussion.

This week's narrative is not one of outright rejection but of cautious optimism. It reflects a journey from fascination and careful endorsement to skepticism and caution—a path characterized by an sincere quest for balance. We advocate for leveraging AI's potential benefits while diligently navigating its ethical and societal implications. This approach aligns with the ethos of Artificial Antics, where we explore technology's frontier with a critical eye, always aiming to be mindful of the broader implications for our collective future.

I may not possess all the answers, but it is crucial that we all understand the intentions behind government and agency use of these tools. Do we want a future where all of our actions are policed by AI systems? An open dialogue and complete transparency regarding system deployment are essential. As we attempt to negotiate this intersection of modern convenience, safety, and freedom, I invite your feedback and thoughts on this complex issue. Please share your perspectives with us on X.com or LinkedIn. Your input is invaluable as we continue to dissect and discuss the evolving landscape of AI and its integration into our daily lives and governance.

Must-Read Articles

Listener's Voice

In this week's Listener's Voice. This week: Xavier asks: "How can I make ChatGPT give me more specific answers to my questions?"?

Great question Xavier! So there a multiple ways to accomplish what you’re looking to do. The first is available to all ChatGPT users, free and paid and is called utilizing the “System Prompt”. When you start a chat, the first prompt you enter is called the System Prompt. ChatGPT places a high emphasis on this prompt and will remember and reference it during the conversation. Knowing this you could use the System Prompt to steer your outputs. Below is an example of using the System Prompt to guide the GPT to act as a social media manager for a popular brand.

Example System Prompt

Imagine that you are a social media manager for a popular brand, and your task is to come up with fresh and engaging content ideas to attract and retain your audience. Your goal is to create content that resonates with your target audience, drives engagement, and ultimately boosts sales. To get started, think about the latest trends, news, and events that are relevant to your brand and audience. Consider what types of content (e.g., images, videos, polls, etc.) have performed well in the past and how you can build on those successes. Now, brainstorm a list of 10 potential content ideas that would capture your audience's attention and generate buzz. Be sure to include a mix of educational, entertaining, and promotional content. Remember to keep your brand's values and personality in mind, and try to create content that feels authentic and unique to your brand.

My first task is to come up with 10 content ideas for social media marketing based on this account description: "Kid's books that tech profound lessons"

This is fantastic if you are bouncing around between multiple types of tasks (job hats) and need the flexibility to tailor each chat based on a persona and goal. Now, if you’ve got the paid version of ChatGPT (Plus, Team or Enterprise), your options open up a bit.

Remember how I mentioned the System Prompt is great for bouncing around between different personas or goals? Well, with the ChatGPT paid accounts you can take that to the next level. This is where Custom GPTs come in!

Use Custom GPTs

Custom GPTs allow you to create that “Social Media Manager” persona forevermore including adding a deeper Knowledge set, and Capabilities the ability to browse the web and the ability to access 3rd party tools via APIs using Actions. More on GPTs in the link below. I also go into more depth in my Tech Deep Dive section further down in this article.

Use Custom Instructions

If you want ChatGPT to have a better idea of who you are and how you would like it to respond by default, you can use Custom Instructions. You can enable your instructions for all new chats

The first text box is What would you like ChatGPT to know about your to provide better responses? Is good to tell ChatGPT about you. Remember this is stuff that you’ll want it to know for every new chat so use sparingly but this can really

The second text box is How would you like ChatGPT to respond? The information in this field let’s ChatGPT know how you like your answers.

Rico and I talk a bit more about Custom Instructions in the clip below:

Xavier, thanks again for your question (you’re not the first to ask us that!). Folks, if you have any follow up questions or want to talk through this more, hit me up at [email protected]!

Catch us on the next episode of Artificial Antics, where we'll explore even further into the technology that's reshaping our world. And hey, don't forget to follow us on X.com/@anticslab for more insights and discussions. See you in the lab!

Mike's Musings

Tech Deep Dive

Mike breaks down a complex AI tool or concept into understandable terms.

Hey everyone, it’s Mike here, doing a deep dive into the differences between deterministic and non-deterministic AI. If you have been with us for a while, you may have heard us speak about the two terms in several episodes of the show. But we figured we would give some better insights via the newsletter here as to the differences between the two. Imagine AI as a vast ocean of possibilities, where some creatures follow a map (deterministic) and others just go wherever the currents take them (non-deterministic). Now, let's sprinkle in some examples to keep things spicy.

Deterministic AI: The Map Followers

Deterministic AI is like a well-trained dog that follows the same path home every day. You give it a command, and it knows exactly how to respond because it's been programmed with specific rules. It's predictable, reliable, and says, "I got this," every time.

Example: Think of a chess AI. You make a move, and it calculates the best possible response based on a massive library of moves and counter-moves. It's like it has a chess grandmaster whispering in its ear, except it's all code and algorithms.

Non-Deterministic AI: The Free Spirits

Non-deterministic AI, on the other hand, is like a jazz musician improvising a solo. It's creative, unpredictable, and sometimes a little out there. This AI uses algorithms that allow it to learn and make decisions based on data, rather than following a strict set of rules. It's the cool cat of the AI world.

Example: Generative AI models like DALL-E 3 and Midjourney are perfect examples. You give DALL-E 3 a prompt, like "a kangaroo in a suit doing a moonwalk," (See below) and it generates an image that's never been seen before. It's not following a specific recipe; it's creating something new by understanding the essence of your request.

The Mix-Masters: Generative AI and Beyond

Now, let's talk about the mix-masters, AI that blends deterministic and non-deterministic approaches to create something unique.

Generative AI (e.g., DALL-E 3, Midjourney): These AIs are primarily non-deterministic. They generate content (images, text, music, etc.) that's original and often surprising. They learn from vast datasets and then apply that learning in unpredictable ways to create new things.

ChatGPT : ChatGPT is like the bridge between the two worlds. When you ask it a question, it generates responses based on patterns and information it has learned from a massive dataset. However, there's a deterministic backbone to how ChatGPT operates, ensuring to follow certain rules and guidelines. So, it is a bit of a hybrid, using both deterministic and non-deterministic elements to chat with you.

Wrapping It Up

In the grand scheme of things, deterministic AI gives us reliability and predictability, while non-deterministic AI offers creativity and adaptability. Together, they're pushing the boundaries of what machines (and AI) can do, from creating art that makes you do a double-take to having conversations that keep you on your toes.

So, whether you're into the precision of a chess-playing robot or the wild creativity of an AI painter, there's something in the AI world for you. And as we continue to explore this ocean, who knows what other fascinating creatures we'll discover?

Feel free to reach out to me if you’d like to talk through anything, you can reach me at [email protected] 

Mike's Favorites

Sharing personal recommendations for technology, AI books, podcasts, or documentaries.

OpenAI Voice Engine

I like the fact that there are a few companies doing voice and video generation well including translation. In my opinion that’s a very short list. If anyone can succeed in releasing solid features around voice and video it’s OpenAI

P.S. Some of my favorite services like HeyGen already use Voice Engine under the hood for their voice translation. Check out my clip below to see how well it works:

Thanks for checking out my favorites section and if you have something you like to share, talk about or ask, hit me up at [email protected]

Contact Us

Got a product, service, or innovation in the AI and tech world you're itching to share? Or perhaps you have a strange, hilarious, or uniquely entertaining experience with AI tools or in the AI space? We at Artificial Antics are always on the lookout for exciting content to feature on our podcast and in our newsletter. If you're ready to share your creation or story with an enthusiastic audience, we're ready to listen! Please reach out with a direct message on X.com or send us an email. Let's explore the possibility of a thrilling collaboration together!

Closing

Elevate your Artificial Antics experience! Subscribe to our YouTube for exclusive AI Bytes, follow us on LinkedIn for the latest AI insights, and share our journey with friends. Don't forget to keep up with us on all major streaming platforms for every podcast episode!

Thank You

As always, a big thanks goes out to our listeners and viewers! For now, keep tinkering, questioning, and laughing – because, in the world of AI, the future is always just around the corner. Catch you on the next wave of Artificial Antics!

Quote of the week: "AI moderation tools in video games highlight the balance between creating safer environments and the risks of over-policing by algorithms." - AI Bytes | AI Policing Video Games​​