Janice R. Andrews
The artificial intelligence tool Magic Media, developed by design giant Canva, is subject to stringent constraints. One such restriction is the prohibition of producing political or medical content. In a recent interview with The Verge, CEO Melanie Perkins stated that the goal of this decision is to stop the spread of offensive or dangerous information.
Avoiding Dangerous Content
Magic Media, an AI feature from Canva, was created with particular content limitations in mind. CEO Melanie Perkins stressed that words related to medicine or politics are ineligible for use with the tool since they may be dangerous or improper. “There are some things we shouldn’t be creating, but Canva was meant to be a place where you could come in and take your idea and turn it into a design,” Perkins said. For example, the program will reject requests to generate photos of political candidates by telling the user, “You can’t do that.”
Content Created by Users Without AI
Users are still able to manually develop graphics on the platform containing political or health-related information, despite these limitations. The restriction solely applies to content created by AI, enabling Canva to preserve the harmony between user creativity and moral content creation.
Terms and Ethical Use of AI Products
Regarding the use of its AI to other sensitive material domains, Canva has also outlined its explicit policies. It is forbidden for the AI to produce contracts, financial or legal advice, spam, or adult content. Canva has also put in place a strict policy against AI scraping, making sure that creators’ content is never used to train the AI without their consent. Users retain control over their intellectual property and can choose at any time not to have their creations used for AI training.
Encouraging Authors and Providing Ethical AI Education
Unless they specifically choose to opt-in, all Canva users are by default opted out of having their private design work utilized for training AI models. Canva established a $200 million fund last year to compensate customers who choose to participate in AI training over the following three years in an effort to better encourage creators. This program demonstrates Canva’s dedication to the moral application of AI and its support of its user base.
Differences between Adobe and Meta
Canva differentiates itself from other industry titans like Adobe and Meta with its approach to AI content creation and training. The creative community has strongly criticized both businesses for their AI strategies. When Meta came under fire last month for training its AI models with public photographs from Facebook and Instagram, numerous artists moved to Cara, a site that completely forbids the usage of AI. Artists fearing that AI might remove their work and content launched a wave of boycotts in response to Adobe’s re-acceptance of its “Terms of Use”. As a result, more people signed up for competing websites like Affinity and Linearity, the latter of which Canva purchased earlier this year.
In an effort to allay these worries, Adobe said in a blog post that user content is their property and won’t be utilized to develop generative AI capabilities. Though they forbid the production of offensive or adult content and caution against relying on AI features for medical advice, the company’s AI standards do not specifically forbid the development of such content in the first place.
Canva’s Promise to Use AI Ethically
Canva’s proactive approach to AI use is indicative of a larger dedication to user protection and ethical content creation. Canva wants to promote a more secure and civilized digital design community by clearly defining what its AI can and cannot produce. In addition to shielding consumers from potentially dangerous information, this strategy promotes the creative community by guaranteeing that their labor is valued and adequately compensated.
