Welcome to The Week in Generative AI, a weekly column for marketers from Quad Insights that quickly sums up need-to-know developments surrounding this rapidly evolving technology.
ChatGPT goes after enterprise customers
OpenAI is releasing a version of its ChatGPT chatbot called ChatGPT Enterprise, which is aimed at large businesses. According to OpenAI, ChatGPT Enterprise offers more security, privacy and higher-speed access to its technology than the consumer version of ChatGPT. Anna Trong of Reuters reports that “early customers [of ChatGPT Enterprise] include Block, Carlyle, and Estee Lauder Companies.”
OpenAI hopes that the launch of ChatGPT Enterprise will make businesses more comfortable with their employees using ChatGPT for tasks at work — already a phenomenon according to a recent Reuters/Ipsos poll (“Some 28% of respondents to the online poll on artificial intelligence between July 11 and 17 said they regularly use ChatGPT at work,” per Reuters).
Trong notes that “Microsoft already offers businesses access to ChatGPT via its Azure OpenAI Service, though in order to use it businesses must be a customer of Azure, Microsoft’s cloud computing platform.”
Writing in CNBC, Hayden Field clarified that “One key differentiator between ChatGPT Enterprise and the consumer-facing version: ChatGPT Enterprise will allow clients to input company data to train and customize ChatGPT for their own industries and use cases, although some of those features aren’t yet available in Monday’s debut. The company also plans to introduce another tier of usage, called ChatGPT Business, for smaller teams, but did not specify a timeline.”
Related coverage:
• “AI Startup Buzz Is Facing a Reality Check” (The Wall Street Journal)
• “What Is ChatGPT Enterprise, What Does It Offer, and How Is It Different?” (Make Use Of)
• “OpenAI seeks to dismiss majority of Sarah Silverman’s and authors’ claims in ChatGPT lawsuits” (VentureBeat)
Google Duet AI joins Workspace apps, plus a new way to watermark AI-generated images
Google this week announced several new artificial intelligence products and partnerships aimed at large businesses, but the headline-grabber is an enterprise-level tool built to watermark and identify images generated with AI. Known as SynthID, the tool embeds an invisible watermark into images created by Google’s text-to-image generator, Imagen. The watermark can be used to identify computer-generated images even if they have been modified.
The other big news from Google is the roll out of Duet AI, its Workspace companion designed to compete with Microsoft’s AI-powered enterprise offerings. “These sprawling do-everything tools are core to the future for both major office suites, and both companies believe AI has the potential to change the way we work,” writes David Pierce in The Verge. Duet AI debuts at $30 per user and promises to enhance apps like Gmail, Meet and Drive. “One way to think about Duet is as a mix of Clippy’s helpful work-ish demeanor and ChatGPT’s creative abilities,” notes Pierce[DSN1] [AM2] .
Related news:
• “Google unveils enterprise AI tools, new AI chip” (Reuters)
• “Google Says Over Half of Generative AI Startups Use Its Cloud” (Bloomberg)
• “General Motors to use Google AI chatbot for its OnStar service” (TechCrunch)
• “Amazon and Google plan to let their voice assistants team up on one device” (Tech Brew)
• “Google’s AI-powered note-taking app is the messy beginning of something great” (The Verge)
• “Google’s new AI-powered search results are ripping off news sites” (Quartz)
Midjourney AI updates
If you’re in the creative field, Midjourney’s new Vary Region feature should be on your radar. It allows you to modify specific portions of an AI-generated image instead of starting anew, somewhat like Adobe’s Generative Fill offering. The utility of this feature is twofold: First, it helps address frequent AI misfires, like anatomically incorrect limbs or fingers. Second, it provides flexibility for iterative design, allowing you to test out new visual elements or expressions on an existing image.
Writing for PCMag, Emily Dreibelbis notes Vary Region’s limitations, pointing out that it’s optimized for changes in large areas of the image, covering 20% to 50% of the canvas, and works best for minor prompt modifications. In other words, it’s not ideal for adding new characters to a scene or for making drastic changes.
Related coverage:
• “A Creator (Me) Made a Masterpiece With A.I.” (The New York Times)
• “Midjourney Takes on Photoshop with Its Own AI Generative Fill-Like Feature” (Beebom)
Further reading
• “The U.S. Regulates Cars, Radio and TV. When Will It Regulate A.I.?” (The New York Times)
• “Meet Aleph Alpha, Europe’s Answer to OpenAI” (Wired)
• “A New Facebook Setting Tells Meta Not to Use Your Data for AI” (Gizmodo)
• “Q&A: Uncovering the labor exploitation that powers AI” (Columbia Journalism Review)
• “Behind the AI boom, an army of overseas workers in ‘digital sweatshops’” (The Washington Post)
• “Growing public concern about the role of artificial intelligence in daily life” (Pew Research Center)
• “Nvidia’s stock closes at record after Google AI partnership” (CNBC)
• “DoorDash launches AI-powered voice ordering technology for restaurants” (TechCrunch)
Thanks for reading. We’ll see you next week.
If you’d like to catch up on prior installments of this column, start by heading to last week’s recap: “The Week in Generative AI: August 25, 2023 edition.”
