OpenAI's Sora From hype to pragmatism

Artificial intelligence (AI) is at the beginning of its journey along the Gartner hype cycle, a conceptual framework that helps track the impact of new technologies over time. So far, AI has not quite reached the end of the first stage –– known as the ‘peak of inflated expectations’ –– when financial markets, enterprises, governments and consumers become overly optimistic about a technology’s potential. Therefore, it may still be some time before it hits the ‘trough of disillusionment’ when those inflated expectations are brought into check by companies failing to achieve what was promised. 

AI text-to-video generators like OpenAI’s Sora are worth placing within the context of the framework because they have been chief in driving AI hype –– especially within video production. Why they have done so can be boiled down to three key areas: efficiency, accessibility, and output:

Efficiency: The instantaneous results produced by these generators collapses multiple stages of the video production into two short steps: writing a prompt and generating a video. Multiple professionals needed for an on-location shoot can be theoretically traded for one person sitting behind a keyboard.

Accessibility: AI video generators enable those with little-to-no time or production skill to start creating videos. They lower the barrier of entry to a proportion of the population who were not video creators before.

Output: They produce film and animation clips that can reach movie-grade results, enabling lower-skilled creators to take on higher-skilled work.

These are all examples of AI video generators working optimally. However, the reality can be mixed. AI visualisations can lead to bogus results. Some of the most impressive videos have come from those with either high-end production experience or have been supported by professional editing and sound mixing teams, which dampens the argument that high-grade video creation is now open to everyone.

The above is all acceptable if AI video generators are perceived as a tool for enhancing video creativity rather than a one-stop-shop process that bypasses most of the production process. The likelihood is that AI video generators will be phased in over time in different parts of the video creation, starting with enterprise and early-stage creators who prioritise speed over creative control, before reaching intermediate and advanced creators when the use case becomes watertight.

Until then, using AI video generators is a tight rope walk. Toys “R” Us, the American toy retailer, released an advert using Sora, which featured an AI representation of its founder Charles Lazarus as a child. The response was divisive, with those impressed by the technology countered by those irritated by its flaws, especially around consistency. In some scenes, Charles appears to shapeshift. Toys “R” Us can take such risks because of the novelty factor surrounding AI. Intrigued consumers are willing to overlook AI’s shortcomings as they are wowed by the possible rather than the results. However, this will soon fade. Then, the onus will be on AI to be indistinguishable from traditional video production –– or risk an audience backlash.

How quickly these shortcomings can be ironed out will dictate the adoption rate for AI video generators among advanced creators. They will struggle if they fail to meet the expectations of an audience conditioned by SVOD services to expect high-budget productions as a quid pro quo for buying a subscription. Where AI can prosper is around tools tailored to specific workflow tasks, such as video upscaling through Topaz Labs, or transcription and subtitles through Subly. These tools integrate AI into video production without overpromising or overreaching.

After capturing the imagination with Sora’s potential, OpenAI should now focus on pragmatism by pinpointing specific use cases that video creators can adopt effectively. By managing expectations in this way, Sora can embed AI video generators into workflows and minimise pullback from users if the trough of disillusionment arrives in the future.

Roles

This report is relevant to the following roles:

Business Strategy