Generative AI’s Impact on Entertainment

This is an excerpt from the white paper, Maximizing Returns: Generative AI’s Impact on Entertainment . This white paper examines the emergence and explosion of GAI and its potential impacts on the entertainment industry including efficiencies and automations within the media supply chain and opportunities to boost revenue.

User Experience

Companies can use GAI to create more positive and engaging experiences by introducing new levels of interaction with brands and content. The result is a user experience that caters to the nuanced preferences of individual users.

Discoverability: Improving UI, Search, Personalized Recommendations

Individually curated content is key to the customer experience, as is an easy search and discovery process. An unhelpful search will lead to frustration at best and drives users to leave the platform in question. GAI can help improve this process because the use of natural language enables a dialog to successively refine results. Traditional search engines work through keywords or phrases that users input into a textbox to answer specific questions. GAI, however, uses conversation to provide dynamic yet comprehensive answers. Companies are seeing the advantage and application to this type of search. ?Amazon recently announced Alexa LLM, which was designed to provide users with real-time information in a conversational format. For example, Alexa can answer successive questions about a museum without needing to provide context after the initial query.

Amazon also recently announced AI-driven improvements to the viewing experience for fans of Thursday Night Football, currently hosted on Prime Video. One new feature is “Defensive Alerts,” which use GAI to identify “players of interest” (those most likely to rush the quarterback) during the game by tracking and analyzing players and plays throughout the game, improving as the game goes on. A LLM operates within a positive feedback loop because it learns from data that is continuously updated. Higher user engagement will feed back into the model to drive better performance and further engagement – and with higher engagement, there is less churn.

Support: Improved Customer Service, Cost Reduction

Exceptional customer relations are a key driver of business success. Companies like Gupshup use GAI to offer a conversational customer support solution that includes subscription renewal reminders, personalized content recommendations, and new show launches.

Digital assistants can also act as customer service agents. Einstein GPT for Service by Salesforce streamlines customer service with always available, conversational support that is personalized to every interaction.

In some cases, if someone has a history with a company, GAI tools can offer solutions to customer needs before the customer even asks. The use of AI for personalization or customer service, however, should be used with caution as efforts can easily backfire if a customer expects a human for a personal matter and realizes the response is artificial.

Generative AI can help automate otherwise time-consuming tasks for both customers and customer service agents. Generative interfaces can offer 24/7 support as well as improve customer service with thorough and polite answers to questions. These tools can provide quick responses that abide by the preferences of each person such as with language and tone. GAI tools can even incorporate a customer’s history with a company to deliver an experience unique to that relationship.

Operational Efficiencies

In already hypercompetitive conditions, entertainment companies are under tremendous pressure to remain vigilant over resources. GAI can provide operational efficiencies that significantly reduce the cost and barriers associated with almost any internal process.

Marketing Materials

GAI can help create personalized marketing materials such as emails, advertisements, videos, and social media content to relieve the pressure of constrained timelines placed on marketing teams. For example,

• Cognitive Mill can work through hours of footage in five minutes to create movie trailers incredibly fast. Tools like these can be used to create and deliver different cuts for different audience segments.
• The AI video personalization platform, Tavus, can take a single recording and generate unique videos for each prospective customer, in minutes. Multi-modal generative outputs can even help pair ad copy with creative images.

Institutional Knowledge: Onboarding and Assisting Staff

GAI tools can also be used within organizations to streamline internal processes with internal knowledge transfer. Companies can digitize the knowledge base of an organization to produce a new employee guide or 24/7 chatbot. Internal use of LLM’s can help efficiently answer questions and quickly draw from private company archives. In a time of increased turnover across all industries, this can help increase productivity for new employees and reduce the support load of senior staff.

• NVIDIA’s platform NeMo can help organizations quickly implement their own secure GAI solution customized to meet specific business needs.
• Levity is one solution that can unburden HR employees and management while a new employee is onboarding at a new company.

Instead of new and current employees searching blindly through company archives or interrupting other staff, an AI assistant can generate a quick answer, drastically reducing hours wasted on administrative tasks. While applicable across all industries, for the entertainment industry, this means an expedited path to understanding the entertainment ecosystem and subject matter expertise. The assistance of GAI could mean a lower learning curve, allowing more employees to use tools previously only specialists would be able to use, to search through movie archive or, to help with creating clips.

Going Global: Cutting Time to Market

Media companies are under constant and increasing pressure to produce and distribute content on a global scale to grow or keep subscribers. Fortunately, GAI can automate time-consuming tasks so that new movies and shows can be brought to global markets faster. GAI can be especially helpful for localization when distributing media in other countries, as well as adding alternate language tracks to videos, including for user-generated content. [BC1]?It can also be used to produce different versions of ad campaigns for different regions, or craft content summaries in different languages so that teams can use the previous iterations from the GAI as a foundation for the final product, expediting the process. The use of GAI can be an enormous cost and time saver and allow media companies to distribute content faster and more easily to new audiences.

Preparing for a Future with GAI: Protecting IP

"The use of these AI tools as it relates to performers' work in projects is a mandatory subject of bargaining under federal labor law," Duncan Crabtree-Ireland, SAG-AFTRA’s national executive director and chief negotiator, recently told The Times. "The law requires [the studios] to bargain over it with us, and that's what we're in the process of doing."

GAI tools scrape data from the web to generate output like text or audio – but that source data was originally created by humans and may have copyright protections. Using GAI to generate new content without the original creator’s consent has both ethical and financial implications.

The use of AI to bypass human talent was a key issue of the 2023 Writers Guild of America (WGA) Strike. The WGA sought to protect creative professionals by placing restrictions on the use of AI for content creation. This included a myriad of concerns around various applications including the ability to impact original and sourced materials, digital clones, and more. [1] The 148-day strike ended with an agreement that requires film and TV producers to obtain actor consent before creating and using replicas of their likeness as well as compensation when those replicas are used.

To date, there are no laws in place to protect against the use of copyrighted material for training an AI model. However, a judge in Washington, DC recently ruled that only art created by humans is subject to copyright protection[2]. Other legislation around training and other GAI uses is still pending. GAI can produce original work with a similar style to the training material, but the output is not usually copied in a literal sense which muddles the boundary between fair use and theft. Examples of recent events:

• A Reddit user fed 32 illustrations by Hollie Mengert into Stable Diffusion’s image generation model to create new illustrations in her likeness, without her permission.
• In February 2023, Getty Images filed for copyright infringement against Stability AI for its Stable Diffusion software, claiming that millions of copyrighted stock images were used to train the AI model without authorization or compensation.[3]

Restrictions around public release of AI-generated content, compensation, and acquired consent from those who contribute to model “training” will be important steps toward proper integration of AI tools within media. Tools like Glaze and Originality.ai are already available for artists and other content owners to protect and prevent unauthorized use of their intellectual property.

Copyright and IP Concerns as a Current Roadblock

Concerns over intellectual property theft will become more commonplace if there are no regulations against the use of copyrighted material for “training” these models. Creatives have made it clear that they will fight against the unauthorized use of their likeness. In the meantime, the use of GAI within the media and entertainment industry will be best implemented in areas such as video editing and post-production to promote efficiency and productivity of human work.

Movement Towards Private LLMs

Companies wishing to implement GAI tools within their organizations may be concerned about the privacy of proprietary assets and knowledge. Tools like ChatGPT ingest queries into its overall model, which can be then used as an output for another user. Proprietary text, images, or video fed into such a model could show up outside of corporate walls. For the entertainment industry, which tightly controls asset distribution and faces billions of dollars lost to piracy each year, asset leakage through a GAI tool is a major vulnerability. Research by Blackberry revealed that 75% of companies worldwide are considering placing a ban on ChatGPT and other publicly available GAI tools. The primary concerns around these tools involve risks to data security and privacy.[4] Fortunately, there are solutions such as Microsoft Azure’s AI platform that allow companies to operate on their own private cloud. From here, companies can spend time building their own custom model with encrypted data.

Companies may also be concerned about the stability and long-term outlook of GAI tools. While ChatGPT created incredible demand and usage for GAI solutions, OpenAI reports spending “roughly $700,000” per day to run the tool and there is speculation the business could go bankrupt in 2024 for losing funding.[5] ?Private LLM tools may prove more stable as the revenue stream is clearer: funding for the work is secured by the end business customer.

Businesses in the entertainment space that augment employee functions with generative AI will benefit from higher productivity and reduced labor expenses. While these tools are famous for increasing efficiency, they do not replace the human touch. Instead, generative AI will complement human creativity. Remaining authentic and using generative AI to enhance human talents rather than replace them will ultimately be the most impactful. Generative AI has the potential to optimize the entire media lifecycle by enhancing the creative process, more deeply engaging viewers, supporting new revenue streams, and creating organizational efficiencies.

This is an excerpt from the white paper, Maximizing Returns: Generative AI’s Impact on Entertainment . This white paper, developed in partnership with FPT Software examines the emergence and explosion of GAI and its potential impacts on the entertainment industry including efficiencies and automations within the media supply chain and opportunities to boost revenue. It highlights concerns around copyright and human rights issues with using GAI, particularly in the media ecosystem.

Finally, it highlights examples of GAI tools already available for use in entertainment and offers guidance on steps media companies can take now to be ready to implement GAI as these models and tools mature, such as building private LLMs, offering 24/7 customer service, focusing on using GAI to enhance rather than replace, and to be mindful of intellectual property protections when creating content.

[1] Actors and writers aren't the only ones worried about AI, new polling shows - Los Angeles Times ( latimes.com )

[2] AI-generated art cannot receive copyrights, US court says | Reuters

[3] Getty Images lawsuit says Stability AI misused photos to train AI | Reuters

[4] 75% of Organizations Worldwide Set to Ban ChatGPT and Generative AI Apps on Work Devices ( blackberry.com )