Last week, we explored how AI is transforming the digital publisher industry, with many publishers seizing the opportunity to implement AI at scale. Building on that discussion, we now delve deeper into the media and entertainment industry insights writ large and shared by Phil Wiser, Paramount Global's Chief Technology Officer, at the recent NAB Show 2024 in Las Vegas.

The emergence of artificial intelligence (AI), particularly generative AI (gen AI), presents both a challenge and a significant opportunity for the media and entertainment industry. Some publishing leaders have moved to seize the moment and implement AI at scale; some remain in the pilot stage, yet others are yet to decide what to do. 

But very few executives talk about it openly.

During a fireside chat with Dan Rayburn at the NAB Show 2024, Phil Wiser, Paramount Global's Chief Technology Officer, outlined his thoughts on artificial intelligence's impact on the media and entertainment industry.

Phil is a senior media executive with over three decades of experience and provides an insider view into one of the world's leading premium content producers. He touched upon the emerging AI trends across various content formats, how they could impact the industry, and ethical and security considerations that media and entertainment professionals should prepare to address.

In this blog, we will examine some of these areas and the strategic steps other media companies can take to boost efficiencies, enhance content, and strengthen engagement and trust.

Billions of dollars back AI.

With billions of dollars in investment, AI will permeate and transform every industry with an expanding array of use cases. Despite lower corporate investment in 2023, investment in Generative AI skyrocketed.

Source: IEEE Spectrum

AI technologies, including large language models (LLMs such as ChatGPT), Natural Language Processing (NLP) algorithms, Machine Learning models, and application layers, on top of all those foundational tools, will become exponentially more capable over time.

Recent McKinsey research estimates that generative AI alone could add between $80 billion and $130 billion to the media industry annually, increasing AI tool's overall impact by 15 to 40%. 

Use cases of AI in the media industry are far-reaching.

According to Stanford University's Artificial Intelligence Index Report 2024, the tech, media, and telecom industries have some of the most robust use cases for AI adoption. Functions like product and service development (44%) top the list, followed by service operations (36%) and marketing and sales (36%).

Stanford University’s Artificial Intelligence Index Report 2024 - Figure 4.4.4

Most media companies are experimenting with AI-powered tools in some way. 

So far this year, NBCUniversal announced a tool that would use AI to automate streaming and linear TV ad buying. Disney has started using AI to serve ads based on content scene and mood, while BuzzFeed intends to become an AI-driven tech and media company. 

BuzzFeed has also fine-tuned LLMs and developed over six chatbots based on its dataset. Its chatbot, Botatouille, uses an AI-based recommendation system to suggest recipes from Tasty, BuzzFeed's food brand.

TMB, home to some of the world's most loved media brands, uses AI to catalog its 95,000+ videos for campaigns and commercials.

The opportunity to use AI in day-to-day operations is immense for efficiency-minded media companies. Advancements in AI tools have ensured they can already automate editorial tasks, aid in research, improve monetization, and even help switch to more engaging formats like video. 

For instance, Aeon's AI-assisted solution can convert text-based articles to high-quality video at scale while retaining its journalistic voice. 

But like most early-stage powerful technologies, AI is a double-edged sword. Gen AI has created concern for some editorial teams, sparking innovation at its best and misinformation, criticism, and legal matters at its worst.

Emerging trends in artificial intelligence across content formats

AI-generated content means tools can produce images in seconds. Songs can be created in the style of singers, dead or alive. More than 3,000 books have been published on Amazon with ChatGPT as co-author, making the idea of writing a book seem possible for many. 

Here's a breakdown of how emerging AI trends across various formants have impacted content creation:

Text

Script Analysis and Modification

Today, AI models can automatically analyze scripts, provide summaries, and suggest edits. They can analyze behavior and preferences to make predictive suggestions on enhancing content styles that resonate with the audience, helping media companies streamline production processes and improve efficiency and engagement. 

Large language models (LLMs) are the main text-handling AIs that power such use cases. While ChatGPT is by far the most popular LLM, many other chatbots and text generators, ranging from Google Gemini and Anthropic's Claude to Meta's Llama-3 and Mistral, are built on top of LLMs.

Media companies can use these native LLM chatbots through prompts (check our handy guide on using prompts) or through software like Grammarly and Hemingway, which enhance user experience and do not require knowledge of prompts. 

Another integrated AI solution is the Microsoft 365 Copilot - an enterprise-ready solution built on OpenAI's GPT-4.

Content Personalization

Content creators and media companies face the challenge of producing personalized, engaging content for a diverse audience overwhelmed by information. AI tools help produce hyper-personalized content based on individual customer behavior, persona, and consumption history. 

For instance, on social media, Gen AI tools enable greater productivity and creativity while blurring the lines between authentic and synthetic.

Image

Material creation/deletion 

Advanced image generation models like DALL-E, Midjourney, and Stable Diffusion have evolved to understand significantly more nuance and detail. With text prompts, they can generate engaging visuals in seconds. (Learn which of these best suits your use case here.)

Earlier this year, Getty Images, whose stock photos are in every corner of the internet, collaborated with Nvidia to launch an AI model trained on Getty's iStock photography and video libraries to generate new licensable images and artwork.

AI is embedded into most image editing tools today. For instance, Adobe Photoshop, which once demanded advanced skills, now uses generative AI to simplify editing. Tools like generative fill analyze pixels to match lighting, color, and perspective, allowing users to quickly fill, edit, or remove image elements while maintaining realism.


Adobe Photoshop (beta) x Adobe Firefly: Announcing Generative Fill | Adobe

Another typical and more visible example is Google's Magic Eraser tool, available in the Photos App. Removing things that get in the way of a perfect photo used to be a sophisticated task, but with the Magic Eraser, AI does it for you in a few taps.

Source: Google

Automated image enhancement

Most image generation models are multimodal. They can receive images as inputs and apply computer vision techniques to automatically enhance quality, adjust lighting, and apply filters for various platform optimization.

This advancement means media companies can use automation in image-editing workflows to output high-quality media content without much manual intervention.

Aeon's Contextual Crop feature is a good example. It uses AI to dynamically crop images (and videos) based on the narrative and visual requirements to ensure your story is always in the spotlight regardless of the format. 

Video

Automated editing

The integration of artificial intelligence in video editing has been one of the most transformative shifts observed in recent years. 

Most of the top video editing tools use AI algorithms. Adobe's Premiere Pro comes packed with features like Generative Extend, Object Addition, and Object Removal, changing the game for film and video creation — speeding up transcription, editing, color, audio, captioning, and delivery workflows. 


Generative AI in Premiere Pro powered by Adobe Firefly | Adobe Video

Similarly, Blackmagic Design's DaVinci Resolve has used deep learning to introduce over 100 new features in its latest update.

More media companies-focused tools like Aeon use AI to curate videos for specific use cases. For example, news organizations can use Aeon as a scalable article-to-video production that:

This is only a glimpse into what's possible. More advanced video AI models like Sora will soon become mainstream and change video generation workflows dramatically. 

Video Quality Control

AI, machine learning, and computer vision algorithms can perform quality checks and identify and resolve real-time issues to maintain high standards.

For example, broadcast content quality can be improved by automating tagging for clip selection and utilizing AI upscaling to transform standard-definition content into higher resolutions.

Real-time analytics and decision-making in broadcasting are being modernized with AI tools that analyze viewer engagement and preferences. These tools enable broadcasters to deliver content more effectively and optimize ad placements.

Mathieu Planche, CEO of Witbe, says AI allows teams to script test scenarios ten times faster and assess video quality with the same criteria as a human viewer.

Related Article: Aeon's Guide to Online Video Players

Audio

Sound optimization

Audio AI is changing how content is created and consumed. According to one estimate, the industry was worth over $4 billion in 2021 and is expected to surpass $14 billion by 2030, growing at a CAGR of almost 16%.

AI products in audio include tools for transforming text into speech, creating voice replicas for dubbing, and powering voice assistants that mimic human tone and cadence. 

Media companies can use these AI tools to enhance audio clarity and balance levels or create soundtracks by learning from existing music styles. They can also accomplish more difficult tasks, such as introducing real emotion in voiceovers when using article-to-video systems:


Aeon’s Real Emotion allows content creators and publishers to fine-tune voices to suit their specific guidelines and preferences.

These AI models understand the content and can adjust parameters based on the nature of the article being converted. This means questions sound like questions. Breaks, pauses, and even laughter (haha!) are generated in response to the script being read.

Voice Interaction

AI-powered voice assistants use natural language processing to understand and respond to user commands, providing an interactive experience. These are increasingly integrated into Internet of Things (IoT) devices to provide personalized experiences and streamline interactions.

The Humane AI Pin and The Rabbit R1 are early-stage products that offer a preview of a future in which personalized AI assistants may someday displace the ubiquitous smartphone.

Rabbit R1 (left) and Humane Ai Pin (right)

Interactive Media

Gaming and VR

AI models can create dynamic game environments that respond to player actions and develop in complexity based on player behavior.

NVIDIA’s Avatar Cloud Engine (ACE) is a suite of technologies that help developers bring digital avatars to life with generative AI. It can voice, animate, and write narratives for video game characters.


Bringing Avatars to Life with NVIDIA Omniverse Avatar Cloud Engine

Virtual reality is also getting a major makeover with AI-powered experiences that offer unprecedented levels of engagement and immersion.

One of the most promising enhancements resulting from the integration of AI in VR environments is the ability to make our interactions with the virtual world and players more intuitive and natural, such as with gesture recognition, voice commands, and other user interfaces. 

These new technologies make the experience more immersive and engaging, which can, in turn, broaden the game's appeal to a broader demographic. 

Augmented Reality (AR)

With the release of Vision Pro, a mixed-reality headset, Apple introduced a new way to interact with the digital world that will shape the future of wearable technology for years to come.

ChatGPT is now available on Vision Pro. 

While there are no immediate use cases, it's intriguing to consider what this means for the future of AI and AR.  Aside from simply speaking and conversing with AI, imagine sending an image straight from your field of vision. 

You might glance at ingredients in your pantry and ask AI to make a recipe using those ingredients, or it might one day explain a difficult problem on your child's homework—think of a beefed-up version of Google Lens.

Ethical and Security Considerations

Generative AI poses both risks and opportunities. As early adopters, media companies will be among the first to tackle its uncertainties, biases, and risks. 

In a recent McKinsey survey of over 100 organizations with over $50 million in revenue, 63% of respondents saw gen AI implementation as a “high” or “very high” priority. Yet 91% of them didn't feel “very prepared” to do so in a responsible manner.

Gen-AI-related risks can be captured in eight main categories:

Source: McKinsey & Company

5 Steps To Pursue AI Innovation with Integrity

To maximize value, media companies must pursue innovation with integrity through select actions. Here are five key steps to take:

1. Establish a control tower to centralize AI innovation, knowledge, and skills

Media companies must move beyond experimentation to help reimagine business models, improve governance, and centralize skills. This will require cross-functional teams to collaborate and identify and prioritize GenAI opportunities while assessing disruptive risks, talent requirements, and data governance needs.

Media companies must identify relevant skills, spot immediate gaps, and train teams in the business and technical aspects of GenAI while leveraging core principles that can inform longer-term decision-making.

Now is a good time to develop a portfolio of targeted GenAI opportunities, revisit the existing catalog of AI use cases, and identify opportunities to incorporate AI where feasible. Pick a healthy mix of “quick wins” and more complex use cases.

2. Reimagine business functions and ways of working

Realizing GenAI’s potential to increase productivity and overhaul business models hinges on new working methods. Organizational structures and workflows must reflect these new ways of working. 

AI will empower employees rather than displace jobs. Invest in the training and development of employees through:

  1. AI Literacy Programs: Educate employees on the basics of AI, including machine learning, deep learning, and natural language processing. This will help them understand how AI can be applied in their roles and how to work effectively with AI.

  2. Customized Training Sessions: Offer training sessions that cater to the specific needs of each department or role. This could include training on AI tools and platforms, data analysis, and how to integrate AI into existing workflows.

  3. Mentorship Programs: Pair employees with experienced professionals with a background in AI. This will provide guidance and support as they learn and apply AI concepts.

  4. Online Courses and Certifications: Provide access to online courses and certifications focusing on AI, including AI ethics.

  5. In-House AI Development: Provide resources and support for employees to develop their AI projects. This will help gain hands-on experience and a deeper understanding.

  6. Industry Conferences and Workshops: These provide opportunities to learn from experts, network with peers, and stay up-to-date on the latest developments in AI. Use networking platforms (such as LinkedIn) to stay updated on emerging trends in AI.

To overcome resistance, media companies can launch small-scale pilot projects using proprietary enterprise data to test GenAI solutions and gather feedback. Results can be used to show how GenAI can enhance existing processes.

3. Put GenAI at the center of your strategy

Media companies that are early to adopt AI will gain experience that could put them at an advantage. This will require prioritizing AI discussions with existing partner ecosystems, highlighting areas of mutual interest and potential cooperation in GenAI. 

Continual monitoring of the AI landscape for new opportunities is critical. Identify new partners – whether start-ups, immediate industry peers, or academic institutions – that can enhance your GenAI initiatives. 

4. Build stakeholder confidence in AI

Some media companies and digital publishers may have already designed ethical frameworks for AI. However, GenAI creates a new set of ethical dilemmas and security risks. 

Media companies should address stakeholder concerns about AI-generated content, such as IP and copyright issues, fake content, security, and data privacy. Employee concerns are no less important. 

Regular dialogue with policymakers is essential as the regulatory landscape evolves, reflecting the impact of AI on society. In the absence of dedicated AI regulation, media companies should prize robust governance to build confidence in their AI applications.

Establish teams and develop policies to implement and supervise ethical AI procedures and ways to monitor and audit them. Before launching LLMs, stress-test them for hallucination, jailbreaking, inappropriate content, or other legal and reputational risks. 

This also includes solutions to track and control content such as deepfake audio, video, or text.

5. Make GenAI integral to your multi-year tech transformation plan

GenAI technologies are evolving at breakneck speed. It’s important to keep your options open while understanding how best to harness them for long-term value creation. GenAI should not be treated in isolation but as an additive to other emerging technology investments.

Summary

Leaders in media and entertainment believe AI will pivotally shape industry trends across a broad range of content formats. However, like every powerful technology, AI is also a double-edged sword. For a lasting competitive advantage, media companies must adopt early and develop robust policies and systems to mitigate risks and uncertainties around the use of AI.