What if we could tell whether a film will be successful before production even began?
According to one AI startup called Scriptbook, we can and all that it takes is a particular AI system that they have developed in-house.
In pitching themselves to industry professionals, they have claimed that their driving algorithm has consistently shown itself to be three times better than traditional methods at figuring out whether a movie will be a box office smash. Reportedly, the key problem with this is that Scriptbook is charging around $5000 for its services when in Hollywood, script reading is sometimes a free service.
A considerable argument can be made, however, that if Scriptbook’s AI system turns out to be all that it is cracked up to be, then it will be well worth the cost.
What’s a $5000, one-time payment as opposed to possible millions lost when a movie goes through the entire production process and then fails to be a hit? The answer seems easy but as we have already suggested, the value of such technology cannot be simply taken at face value. Even in the face of this fact, Scriptbook has reportedly directly claimed that it could have prevented Sony Pictures from releasing 22 movies that ended up costing them millions. To add to this, Scriptbook also said that their AI can predict the future rating of films as well as most details about characters like the emotions they will show in certain scenes.
Finally, as if these claims are not ambitious enough, they even went so far as to assert that their AI’s foundational algorithm is able to predict the target audience of films. In saying all of this, it is important to remember that the evidence to back up such a claim is shaky at best.
It seems that any assertions made to the effect that Scriptbook’s AI has achieved such success, fall flat due to the fact that they have kept any technical specifications close to the breast. If Scriptbook does in fact publish at least some part of these specifications, then the whole narrative might change. In the end, any chance of Scriptbook achieving measurable success could be said to depend on decentralization in the context that it can be understood from the Blockchain world.
The lack of decentralization related to technical information on the project is not the only issue that rises to the surface. There is also the question of how their bold claims on their AI’s success rate can be reconciled with the generally accepted fact that a significant gap exists between AI and humans in every vocation. AIs simply cannot keep up in terms of emotional factors involved in decision making. With this in mind, we can safely say that the time when AI cuts humans out of careers that require a lot of emotional decision making has never been further away. In this specific case, script reading does fit this mold. Experts on the subject need to reasonably sense how the actors and the audience will feel about the film.
In Scriptbook’s case, however, they do not seem to have the aim of ever eliminating the need for humans in this vocation. On the contrary, they have said that their AI is meant to augment the abilities of script readers as well as film executives and actually improve the overall experience of these professionals, in connection with improving the bottom line.
So, all in all, if we take this overarching goal into consideration, in place of the idea that they want to replace human workers, the primary use case of their tech becomes at least a bit more believable.
It should be made clear that in doing so, we have to also understand that if they cannot prove the altruism in their AI with regards to human workers, than their efforts might fail entirely.
Resources:
Primary Source: https://variety.com/2018/artisans/news/artificial-intelligence-hollywood-1202865540/
