Adobe Inc. Chief Product Officer Scott Belsky said the company has never trained its generative artificial-intelligence services on customer projects, responding to a wave of user criticism.
(Bloomberg) — Adobe Inc. Chief Product Officer Scott Belsky said the company has never trained its generative artificial-intelligence services on customer projects, responding to a wave of user criticism.
“We have never, ever used anything in our storage to train a generative AI model,” Belsky said Tuesday in an interview with Bloomberg. “Not once.”
Artificial intelligence products like OpenAI’s Dall-E are called generative AI because they take in content to generate new material. But the technology has ignited anxiety among artists who fear it may use their work without compensation. In early January, a screenshot of Adobe’s terms of service was spread on social media and described by some as the company using customer images or videos to train multimedia generation AI models.
Belsky said the decade-old policy is in place to allow products to be analyzed to improve features, not for image generation. “We are rolling out a new evolution of this policy that is more specific,” Belsky said. “If we ever allow people to opt in for generative AI specifically, we need to call it out and explain how we’re using it.”
For more: OpenAI, GitHub AI Tools Draw Legal Scrutiny Over Fair Use
The critical reaction was a “wake-up call,” Belsky said. “We have to be very explicit about these things.”
In the rapidly developing AI imagery space, Adobe hopes to differentiate its generative tools by integrating them into existing creative software. The maker of Photoshop and Illustrator highlighted many new AI-powered tools at an October user conference. For media created with AI hosted on Adobe’s stock photo platform, the company requires images to be labeled and abide by restrictions such as a ban on identifiable people.
More stories like this are available on bloomberg.com
©2023 Bloomberg L.P.