Welcome and happy Tuesday!

Today is jam-packed with exciting AI updates. OpenAI is launching ChatGPT Enterprise, and there's a new AI art tool in town. Also, we’ll delve into the potential dark side of voice cloning AI, and explore Google's innovative tool for responsible AI image creation. Let's dive in!

IN TODAY’S ISSUE

• 🧠 OpenAI launches ChatGPT Enterprise

• 🎨 Move over Midjourney, there's a new AI artist in town: Ideogram.ai

• 🥸 The new voice of deception: the dark side of voice cloning AI

• 🦾 Google's new tool to tag AI images with invisible watermarks

• 💡 3 quick and fresh AI updates

• 🎨 AI art of the day

Read time: 4 minutes

CHATGPT

🧠 OpenAI launches ChatGPT Enterprise

OpenAI has introduced ChatGPT Enterprise, a version of ChatGPT that is geared towards organizations, offering improved security, customization, and unrestricted, high-speed access to the robust GPT-4 model.

• ChatGPT Enterprise ensures business-level privacy controls and encryption, and underlines that no data will be leveraged to train OpenAI's models.

• The enhanced version provides unrestricted speedy access to GPT-4, without any usage limitations, and features a 32k expanded context window (compared to approximately 8k in the standard model) for larger files.

• It encompasses a rebranded version of the Code Interpreter, termed 'advanced data analysis', and includes shareable chat templates to facilitate collaboration and establish uniform workflows.

With features like unlimited access to ChatGPT-4, up to two times higher speed, four times longer context windows, and the ability for teams to share useful chats and prompts, it not only tackles the privacy issues but also enhances efficiency. This is substantiated by Asana's report, which states that ChatGPT Enterprise has already reduced research time by an average of one hour daily. Being SOC 2 compliant, ChatGPT Enterprise addresses the privacy concerns of major corporations that have previously prohibited the use of ChatGPT at work.

AI TOOL OF THE DAY

🎨 Move over Midjourney, there's a new AI artist in town: Ideogram.ai

There's a new and fierce competitor ready to challenge the current leaders in the world of generative text-to-image AI: Ideogram.ai.
In a somewhat surprising move, it opened its doors for free access to everyone yesterday, following an invitation-only beta phase.

We took Ideogram for a spin and found it to be impressively mature and user-friendly, even though the image quality doesn't quite match up to that of Midjourney yet.

The unique feature that distinguishes Ideogram from its rivals is its capability to generate images complete with readable text. This opens up a plethora of opportunities that are simply not possible with Midjourney.

As of now, Ideogram is free of charge and incredibly quick at generating images.

Given this strong start, we are eagerly anticipating the future advancements of this tool.

NEWS

🥸 The new voice of deception: the dark side of voice cloning AI

The world has entered a new era of deception with the rise of audio deepfakes, a technology that uses artificial intelligence to create convincing voice clones.
This technology is now being utilized by scammers to deceive people into thinking they are speaking with a loved one or a trusted figure, leading them to part with their money or confidential information.

Experts warn that this technology can be used for misinformation and scamming, posing a significant threat to individuals and society. Despite efforts from companies to increase service prices and manually verify new accounts, experts believe that legislation should be put in place to prevent such scams in the future.

To avoid being scammed, it is recommended to be mindful of unexpected and urgent calls, verify the caller's identity by asking personal questions or using a different communication channel, and taking a step back to assess the situation logically.

AI IMAGES

🦾 Google's new tool to tag AI images with invisible watermarks

Google Cloud, in partnership with Imagen, is beta testing SynthID, a tool that embeds an imperceptible digital watermark into the pixels of AI-generated images for identification.

This innovation is crucial in an age where generative AI technologies create realistic 'synthetic imagery', raising the risks of spreading false information. Traditional watermarks can be edited out and do not provide a robust solution.
SynthID, however, remains detectable even after common image modifications, enabling responsible creation and identification of AI-generated images.

This tool, developed by Google DeepMind and refined with Google Research, is part of a broader effort to build AI-generated content responsibly and uphold trust between creators and users.

While SynthID is not foolproof against extreme manipulations, it provides a promising approach for responsible AI-generated content handling and could evolve for use across other AI models, mediums, and Google products in the near future.

QUICK AND FRESH

🖥️ Bing struggles to gain search market share against Google despite big AI push

Despite Microsoft's ongoing efforts to promote the advantages of its Bing AI, the latest statistics reveal that the company's portion of the search market remains stagnant at approximately 3%. (link)

🍄 Perilous picks: AI-crafted mushroom foraging books proliferate on Amazon

Experts are expressing concerns that books created by ChatGPT, which are available for purchase on Amazon and aimed at novice foragers, might inadvertently lead to fatal consequences. (link)

🍲 DoorDash introduces AI voice ordering solution

Voice ordering technology will enable select operators to increase revenue and maintain high-quality customer experiences, without increasing labor costs or sacrificing hospitality. (link)

AI ART OF THE DAY

Minty Fresh

FEEDBACK

Until next time, stay informed!