Responsible AI: Promoting Diversity in the Film Industry
Examining modern applications of AI and evaluating the role human influence has on its success
“With great power comes great responsibility.”
As movie trends go, most of us have been exposed to this phrase. It has a certain pull that captures our imaginations and conveys an inspiring sentiment, appealing to a very human side of the audience.
How, then, can we apply it to something that isn’t human?
The public image of AI
One of the definitions of responsibility that fits this context is: “a moral obligation to behave correctly toward or in respect of.” Echoing the message of the movies, this definition suggests that though ability and proficiency come with potential, they require guiding principles to ensure that potential is realized in enabling and empowering ways. Given the mystique and enigma surrounding it, artificial intelligence (AI) happens to be one of those abilities.
In work-related conversations, I often field the question: “Do you believe that AI poses a risk to human civilization?” I understand the origin of this query — AI resistance and uprising is a popular theme in science fiction. Between Terminator, Battlestar Galactica, Mass Effect, and I, Robot, we have detailed descriptions of scenarios where AI takes an antagonistic stance toward sentient species. The usual concerns surround the way AI is portrayed in film, possessing traits like omniscience and superhuman cognition to attempt its grand, malicious plans. That, coupled with recent instances of AI failures, seems to validate the concern that AI development and research may not be in humanity’s best interest.
I like to see the actions of an AI entity as a reflection of its interactions. In many ways, AI resembles nascent intelligence, one similar to that of a child. Children learn from observation and emulation — they experience the world around them and find their way by echoing the behaviors of others. Development of AI is similar. Like how a child learns to recognize dogs from cats, AI tries to differentiate between photos of dogs and cats by seeking feedback on its guesses, developing identification cues in the process.
This analogy can also extend to some of the more infamous examples of AI-related events — a prime example being Microsoft’s Tay. In 2016, Microsoft released the “teen” AI bot in a public experiment, intending it to be a conversational companion for a young crowd. However, it had to be taken down within hours of launch after it started tweeting some inflammatory sentiments.
Expert opinions suggest that Tay learned from the users it interacted with — given the nature of chatbots — and started echoing the themes of those conversations. In other words, it’s theorized that Tay learned those behaviors from the people it spoke to, in the same way children use their limited experiences to form judgments, parroting words they hear from sources around them. Just as it is the responsibility of parents, guardians, teachers, and other sources of influence to instill socially appropriate behaviors in children, for AI this responsibility falls to the people designing the system.
AI in media and entertainment
One of the greatest sources of influence in today’s world is the entertainment industry. As a global society, we consume more than $2 trillion of media and entertainment annually — a number that’s projected to only increase by 2025. Movies, TV series, and video games can form a social bridge between people through shared preferences, opinions, and theories. Fandoms have a way of uniting people from different geographies and communities, and the impact of media on AI is profound.
AI entities are capable of learning from people and, at the same time, paving a novel path to reach the same destination. Some research labs use games as a way of training AI, and there are various examples of AI entities playing against professional gamers to help the players discover new winning strategies. AI is also starting to be used on movie script development and has progressed from writing mostly gibberish nonsense from a few years ago, to a fairly coherent and complicated script. AI is gradually becoming more flexible, creative, and sensitive to human interactions.
Slalom was recently working with the diversity, equity, and inclusivity (DEI) division of a movie studio to identify opportunities for how AI can help promote diversity in the organization, both in front of and behind the camera. While the principles of diversity encourage celebrating a variety of ideologies and perspectives, AI learns by making generalizations based on its experiences. This makes applying AI to DEI incredibly difficult since diversity is a challenging concept to generalize — for AI, understanding nuanced behaviors is almost counterintuitive.
Understanding the comfort of team members in a work environment is key to defining a positive and inclusive experience. To accurately identify and capture this data, we can use AI to gather and evaluate crew feedback, as well as recommend solutions that could meaningfully contribute to the promotion of diversity and inclusivity on set.
The idea behind this initiative was to identify areas of concern, recognize which of them needed the highest level of attention, and determine what could be done. This initiative would require a combination of natural language processing (NLP) techniques that can work together to identify the sentiments, themes, and topics in survey responses. By codifying this in an algorithm, we can process survey responses with better-than-average accuracy while processing a larger volume of responses than humanly feasible.
Once the solution determines common themes and topics, each point will be prioritized based on how often it appears in survey responses, the rating that was associated with the response, and how long the concern had been outstanding. Once the most pressing concerns are identified, the tool would then examine the standard policy for addressing those concerns and make recommendations accordingly.
While the implementation of this solution is currently in the preparatory phase, it will be able to understand, evaluate, and act on the crew’s needs. Through this AI entity, we’ll be able to promote the interests of those on set who would otherwise likely be overlooked. In turn, this has the potential to create a rippling effect of promoting DEI throughout the industry, both behind and in front of the camera.
The future of AI
I’ve personally witnessed how AI solutions can have a massive, positive impact by promoting personal and professional growth, whether it be fiscal, cultural, or social. Last year, in the drive to combat unemployment, AI was used to identify potential employment tracks for those looking to advance in their roles. The AI would recommend which roles were immediately available to them and which additional skills would be most beneficial, depending on the type of role that person was targeting. In my time at Slalom, I’ve seen similar examples in the education sector where AI has improved infrastructure, enabled more real-time insights, and assisted with cybersecurity.
Through AI, we’re creating, shaping, and molding an entity devoted to a specific function — and it’s tireless, meticulous, and thorough in its pursuit of that function. While AI might wield a lot of power, it’s in need of a guiding force, and we have the responsibility to shape a better future by channeling the positive potential of AI.
Slalom is going on tour with AWS, Databricks, and Tableau to help your organization harness the power of AI responsibly. Secure your spot now.
+++
Slalom is a global consulting firm that helps people and organizations dream bigger, move faster, and build better tomorrows for all. Learn more and reach out today.
