Welcome to The Week in Generative AI, a weekly column for marketers from Quad Insights that quickly sums up need-to-know developments surrounding the rapidly evolving technology.
Generative AI looms large at Cannes Lions
At the Cannes Lions International Festival of Creativity, held this week in the South of France, AI had a major presence, with over 200 AI-related events and activations, as Pete Blackshaw, CEO of venture fund Cintrifuse, notes in an Ad Age guest post.
Blackshaw engaged in a little experiment by enlisting ChatGPT and Google Bard to pick Cannes Lions winners over the years. They both ended up highlighting timeless campaigns such as Apple’s “Think Different” and Volkswagen’s “Think Small.” In the PR Lion category, AI chose transformative campaigns including the viral “Ice Bucket Challenge” and “Like a Girl” by P&G’s Always.
• “It’s OK to talk about AI now: Publicis hits back at haters by celebrating Marcel in Cannes” (Adweek)
• “Cannes Today — AI obsession, Edelman stalked, Spike Lee’s Cannes take and more” (Ad Age)
• “Google and Omnicon strike generative AI deal to bring text and image creation to advertising” (Ad Age)
Adobe is looking to quell creator concerns
Adobe’s new AI tool, Firefly, has sparked controversy. Creators on the Adobe Stock platform, where Firefly’s AI was trained (in part) on their images, claim this was done without clear consent. VentureBeat’s Sharon Goldman interviewed creators and found that they’re worried Firefly’s ability to generate images will reduce demand for their work.
She reports that creators argue it’s unethical for Adobe to use their images to train an AI that competes with them, despite legally agreeing in the terms of service. “The artists say that even if Adobe did not do anything illegal and this was indeed within their rights,” Goldman writes, “the ethical thing to do would have been to pre-notify their Adobe Stock artists about the Firefly AI training and offer them an opt-out option right from the beginning.” Creators also criticize Adobe’s lack of transparency regarding their images’ use.
Adobe has responded by emphasizing Firefly’s beta status, stating that it’s committed to compensating creators and actively seeking feedback.
Earlier related coverage:
• “Adobe is so confident its Firefly generative AI won’t breach copyright that it’ll cover your legal bills” (Fast Company)
• “Adobe to Sell Generative AI Subscription With Copyright Assurance” (Bloomberg)
Meta would like to speak with you
Meta’s Voicebox is a new generative AI tool — not yet publicly released — that produces audio. It has the capability to synthesize speech by training on a surprisingly brief clip. Oscar Gonzalez at CNET explains that Voicebox uses a 2-second audio sample and “with that clip, it can match the audio style as well as do text-to-speech generation or re-create a portion of the speech that may have been interrupted by some external noise.”
The potential applications of this technology are vast, ranging from giving a natural voice to virtual assistants or nonplayer characters in the metaverse, to aiding visually impaired individuals by having messages read out in familiar voices.
• “You can’t use Meta’s Voicebox AI — it’s too dangerous” (Laptop Mag)
• “Meta launches ‘Human-like’ AI image generation model” (PetaPixel)
• “Should You Start a Generative AI Company?” (Harvard Business Review)
• “Banking and finance companies hop on the generative AI bandwagon” (Tech Brew)
• “73% of consumers trust what generative AI wants us to see” (ZDNET)
• “Microsoft’s Satya Nadella is betting everything on AI” (Wired)
• “The Impact of Generative AI on Hollywood and Entertainment” (MIT Sloan)
Thank you for checking in on our continuing coverage of the generative AI beat. See you next Friday.
Previously: “The Week in Generative AI: June 16, 2023 edition”