SKF Group, a Swedish manufacturing and bearings company, deployed artificial intelligence to resurrect the likeness of Greta Garbo, the legendary Swedish American actress who died in 1990, for a recent advertisement. The AI-generated replica aimed to capitalize on Garbo's iconic status and Swedish heritage to promote the company's products.
The campaign reflects a widening trend among corporations seeking to leverage deepfake technology and generative AI for marketing purposes. Brands increasingly view synthetic celebrity likenesses as a cost-effective alternative to licensing fees, negotiating with estates, or dealing with the unpredictability of living endorsers. For SKF Group, the choice of Garbo offered a distinctly Swedish angle in a domestic market, tapping into national pride and cultural nostalgia.
However, the execution faced immediate credibility problems. Not all viewers found the AI replica convincing, with observers noting that the synthetic rendering fell short of delivering photorealistic authenticity. The uncanny valley effect, where AI approximations of humans appear unsettling rather than compelling, appears to have hampered the ad's effectiveness. This underscores a persistent technical limitation in generative AI: while the technology has advanced rapidly, creating flawless human likenesses remains difficult.
The campaign also raises thorny legal and ethical questions. Garbo's estate did not authorize the use of her likeness, raising concerns about publicity rights and whether consent remains necessary for deceased individuals whose images enter the AI training pipeline. Swedish law, like many European jurisdictions, does provide protections for personality rights and image use even after death, though enforcement against AI-generated content remains untested in courts.
The incident arrives amid broader regulatory scrutiny of AI in advertising. The European Union's AI Act imposes restrictions on high-risk AI applications, including those generating synthetic media. These rules could require transparency disclosures when advertisements feature AI-generated likenesses. The United States has begun exploring similar safeguards through the Federal Trade Commission, which has flagged deepfakes and synthetic endorsements as potential deceptive practices.
For marketing professionals and corporate strategists, the SKF campaign illustrates both the appeal and peril of AI-generated celebrity endorsements. While production costs drop sharply using synthetic likenesses, reputational risk rises. Audiences increasingly expect transparency about AI involvement in advertisements. Failed executions damage brand trust more than traditional advertising missteps, since they signal either technological incompetence or ethical corner-cutting.
The broader implication concerns the future of celebrity endorsements and talent licensing. As AI improves, studios and talent agencies will face pressure to renegotiate their business models. Estate holders of deceased celebrities may demand licensing agreements even for AI reproductions. Regulatory frameworks will likely require explicit consent and clear disclosure of synthetic media in advertisements within the next two to three years.
SKF Group's misstep with Garbo demonstrates that AI capability does not equal marketing wisdom. The company bet on nostalgia and cost savings but miscalculated audience perception and legal exposure. As deepfake technology matures, companies deploying synthetic likenesses without clear consent or regulatory compliance will face escalating reputational and legal consequences.