Google Fined $270 Million in France 💸

Welcome to another edition of Horizon AI,

Google can't seem to catch a break. The company, whose AI offerings sparked controversy just weeks ago, is now facing a hefty fine for copyright infringement in France.

Let's get into it!

Read Time: 4.5 min

Here's what's new today in the Horizon AI

• Google Hit with $270 Million Fine Over AI Training Practices 💸

• OpenAI's GPT Store Grappling With Influx of Spam

• AI Research: Researchers Gave AI an 'Inner Monologue'

• AI Tutorial: How to clone your voice with AI

• AI Tools to check out

• The Latest in AI and Tech 💡

AI News

GOOGLE

Google Hit with $270 Million Fine Over AI Training Practices 💸

Google was slapped with a hefty €250 million (around $270 million) fine by French regulators on Wednesday, stemming from how it trained its AI chatbot Bard, now rebranded as Gemini.

Details:

• Regulators claim Google violated commitments from a previous settlement about negotiating deals with French media outlets for their content.

• Allegations state Google used journalists' work to train Bard/Gemini without informing the publishers.

• In a statement released Wednesday, Google said the fine was "not proportionate" to the allegations but agreed to pay because it was "time to move on."

The training methods employed by tech companies for their chatbots continue to be a controversial topic, one that has already been brought up in court on numerous occasions.

Source

TOGETHER WITH MAXAI.ME

MaxAI.me - Do more Faster with 1-Click AI

MaxAI.me lets you chat with GPT-4, Claude 3, Gemini 1.5. You can also perfect your writing anywhere, save 90% of your reading & watching time with AI summary, and reply 10x faster on email & social media.

Join 1M+ users on MaxAI.me to do more faster!

OPENAI

OpenAI's GPT Store Grappling With Influx of Spam

OpenAI's official GPT Store appears to be facing growing pains as it rapidly expands, with a flood of bizarre and potentially copyright-infringing AI models slipping through the cracks.

Details:

• Numerous GPT models claim to generate art, stories, or characters styled after popular Disney, Marvel, and other major media franchises without proper authorization.

• Many listed GPTs advertise their ability to bypass academic plagiarism detection tools, violating OpenAI's rules against promoting academic dishonesty.

• The store features GPT personas imitating famous individuals like Elon Musk, Donald Trump, and celebrities without their consent.

While the GPT Store has experienced rapid growth, with over 3 million GPTs registered, this explosive expansion appears to have surpassed OpenAI's content moderation capabilities.

Source

AI RESEARCH

Researchers Gave AI an 'Inner Monologue'

Researchers from Stanford University and Notbad AI Inc. have developed a new algorithm called Quiet-STaR that gives AI assistants the ability to mull over multiple possible responses before providing their final answer.

Details:

• Current chatbots simply generate responses directly from their training data, without stopping to evaluate multiple options

• Quiet-STaR directs chatbots to first produce several candidate responses to a query, then compares them to the original question to select the most appropriate one.

• The algorithm can also learn and improve its response selection capabilities over time.

• When integrated with the open-source Mistral 7B chatbot, Quiet-STaR boosted its score on a standard reasoning test from 36.3% to 47.2%. Math performance also improved significantly.

The researchers believe Quiet-STaR could allow existing AI assistants to provide more accurate, thoughtful responses overall. It is an important step towards making interactions with chatbots feel more natural and human-like.

Source

AI Tutorial

How to clone your voice with AI

Here’s how you can clone your voice with AI and use it to create narrations with your voice:

1. Go to MyVocal.ai website and Sign in

2. Click on the Voice Clone section located in the left column

3. Click on the Record Samples

You will see that a list of phrases appears, which are the ones you have to say. In this process, you'll need to record 25 voice samples. All you have to do is click on the record button and say what appears in a text.

4. Click on the Submit for Clone Training

You will send these samples to train the AI.

5. Wait for the Create TTS option to appear under Text-to-Speech in green.

Until it appears, what you'll see is Processing, which tells you that your voice is being processed.

And that’s it, your AI voice is cloned.

6. Listen to the AI speak with your voice.

Click on the Text-to-Speech section in the left column. Then, you simply need to type the text you want and click on Generate. When you do this, make sure your AI voice is selected at the top to use it.

AI Tools to check out

🎨 Pixelicious: AI-powered online tool that converts your photos into retro pixel art with one click.

👨‍💻 Safurai: AI code Assistant that saves you time in changing, optimizing, and searching code.

▶️ Thumbnail AI: Use AI to analyze and optimize your YouTube thumbnails. Get higher click-through rates and maximize your video's visibility.

🎧 Podwise: It transcribes, extracts insights, and summarizes dense podcast episodes.

🖼 Scribble Diffusion: Turn your sketch into a refined image using AI.

The latest in AI and Tech

Tennessee becomes first US state with law protecting musicians from AI

Tennessee has become the first U.S. state to enact legislation safeguarding musicians and other artists against unauthorized AI impersonation. The Ensuring Likeness Voice and Image Security Act, or 'ELVIS Act,' fortifies the state's existing right of publicity law with robust AI-specific protections.

University of South Florida launching college for AI

The University of South Florida has unveiled plans to launch a college focused on artificial intelligence, cybersecurity, and computing. The college will be the first of its kind in Florida and one of the earliest nationwide.

Jensen Huang: AI hallucinations are solvable, and AGI is 5 years away

Regarding AI hallucinations, Huang expressed confidence in their solvability by making sure the AI does research before answering. As for artificial general intelligence, he proposed achievable benchmarks that could be met within 5 years.

Apple Explores AI Partnership with Baidu

Apple has engaged in talks with Baidu, the Chinese tech giant, regarding the potential integration of Baidu's generative AI technology into Apple's devices within the country. This partnership could prove instrumental in ensuring compliance with China’s AI regulations.

Copilot gets its own key on Microsoft's new Surface devices

Microsoft has unveiled the Surface Pro 10 for Business and Surface Laptop 6 for Business, calling them “the first Surface AI PCs built exclusively for business.” A key highlight is the inclusion of a dedicated Copilot button on the keyboard for quickly accessing Microsoft's AI chatbot assistant

That’s a wrap!

Thanks for sticking with us to the end! Let’s stay connected on LinkedIn and Twitter.

Not subscribed yet? Sign up here and send it to a colleague or friend!

See you in our next edition!

Gina 👩🏻‍💻