TVN’S MANAGING MEDIA BY MARY COLLINS



Mary Collins

One look at the schedule for the 2024 NAB Show and you might be convinced that artificial intelligence (AI) offers the solution for all that troubles today’s media businesses. Planned sessions tout AI answers for operations, production and everything in between.

On the business side, attendees can select from pages of offerings including those dedicated to business models; strategies for sales, marketing and operations; and how AI can be used to support diversity and inclusion initiatives.

If you’re more interested in the content side of the business, consider sessions devoted to AI for the newsroom; how artificial intelligence tools can be used to produce reality television or documentaries; and uses to create storytelling efficiencies or personalize content for audiences.

The exhibit floor will also be packed with AI solutions. There are multiple designated areas where this new technology is the focus, and at least three of the scheduled tours will help attendees see the AI opportunities on offer.

While garnering a lot of attention right now, AI programs have been around for some time. Media businesses are already benefiting from voice-to-text solutions including real-time closed captioning and content transcription. Some streaming businesses offer AI-enabled content suggestions based on consumers’ prior viewing. Even business departments have jumped on the bandwagon using bots to automate time-consuming routine tasks.

BRAND CONNECTIONS

What has changed is the debut of generative AI (gen AI) programs including Open AI’s ChatGPT along with offerings from scores of other companies including Google, Microsoft, Dalle-E2, Midjourney, Anthropic, IBM and others in a continually growing list. These have been fed vast quantities of content and are programmed to provide outputs that seem to be new instances of text, sound or visual images.

The Possibilities

The creative applications and business opportunities for media companies adopting this new technology are seemingly endless. Consider the ability to dub or subtitle content in minutes. Even better, it’s possible to use voices that sound like those of the original actors. How about being to personalize content by recreating whole scenes to comply with varying country regulations or even local mores?

Do you remember the November 2004 controversy related to local television stations refusing to show the movie Saving Private Ryan? Those stations were concerned about the use of profanity, particularly the F-word, that could result in an FCC complaint. Today’s generative AI could easily replace that word with an innocuous “fudge,” matching each actor’s voice and intonation.

Station groups are also looking for cost-effective content options. With GenAI’s ability to virtually construct actors, scenes, sounds, voices and even storylines, is it unreasonable to imagine a local series written, produced and directed by AI? This is not a new idea. There was that short-lived low-resolution animated Seinfeld parody called Nothing, Forever that continuously updated on streamer Twitch. (It was suspended just over a year ago after the main character went off the rails with a transphobic rant.)

One of the planned speakers for an NAB session, Soul Machine’s Fay Wells, seems to be foreshadowing this use case in her quote for the guide to the show’s AI sessions saying: “The democratization of AI tools has made it easier than ever to not only curate highly personalized content and sophisticated interactional capabilities, but also produce, scale and monetize this work at levels once inaccessible to those outside of the multi-million dollar studios.”

If you are tempted to dismiss Wells’ observation as mere sales hype, you need only read the widely-quoted Hollywood Reporter story about Tyler Perry’s pause of a planned $800 million studio expansion in Atlanta after seeing an early demo of OpenAI’s Sora. Perry explains, “actually seeing the capabilities, it was mind-blowing.”

On the business side, use cases for gen AI include harnessing its ability to analyze vast quantities of data and identify patterns. I’ve already mentioned how this is being used to recommend content. Business departments may also deploy the software to summarize contracts and agreements. These summaries can be useful in calculating return on investment (ROI) for specific terms and conditions as well as for flagging all manner of problems and opportunities. Software optimized for consumer targeting can improve advertising effectiveness. Additionally, generative AI’s proven ability to write code can both speed up software customization and reduce the cost of such enhancements.

The Potential Problems

The addition of AI, particularly generative AI, can also introduce significant risks, a lot of it financial risk. Let’s start with ROI calculations. According to a recent article in the Wall Street Journal, they might not be as straightforward as one would think. The authors explain that companies are introducing calculation variables including productivity, employee satisfaction and even communication with customers. The problem becomes even more complex when factoring in costs for customization, comprehensive employee training and ongoing maintenance.

What about the long-term viability of the software itself? Gen AI is a new technology with lots of companies offering solutions for media businesses. Not all of them are going to survive. Just as it did with traffic software and other technology solutions, the industry will ultimately come to rely upon a few providers. Companies that opted to use something other than one of the surviving solutions will ultimately face the costs (and headaches) associated with transitioning to another platform.

There are other risks as well. As I explained in an earlier column, data is a big one. The first thing to consider is that consumers are becoming increasingly protective of their privacy. Governments are responding with legislation intended to protect that privacy. Penalties for violations can be severe. Additionally, less accurate data could limit the value of targeting and customization efforts.

Underlying data reliability and accuracy is also a concern. Most currently available gen AI providers have been circumspect, at best, regarding the data used to train their programs. Additional tweaks meant to fix identified shortcomings can lead to inaccurate outcomes. One well-reported example of this is the Google Gemini depiction of U.S.’s Founding Fathers as including Black men.

AI has evolved at a speed never anticipated by copyright law. There are a number of lawsuits claiming copyrighted works were used without permission. Companies such as OpenAI argue that they only used data available under so-called “fair use” provisions. It will be years, if ever, before we learn whether courts agree with that characterization. Users need to be aware that any data used to query to public programs, such as Microsoft’s Copilot, becomes part of the program and will be shared indiscriminately, regardless of copyright.

On the other hand, current court rulings say that the output of gen AI programs cannot be copyrighted. That means that compelling AI-generated treatments or characters can be used by others without obligation.

Finally, news outlets that pride themselves on their credibility will need to take particular care to ensure that they don’t mistakenly report AI-generated falsehoods. CBS network recently announced the hiring of an executive director for its fact-checking news unit called News Confirmed. The group’s charge is to identify misinformation and deepfakes. It will be interesting to see how others tackle the problem.

As the focus on AI at the upcoming NAB Show (and January’s CES before it) clearly shows, AI is something media companies ignore at their peril. Given that there are at least as many downsides as upsides, savvy media professionals will take opportunities to learn all they can before deciding the best path forward for their businesses.


Former president and CEO of the Media Financial Management Association and its BCCA subsidiary, Mary M. Collins is a change agent, entrepreneur and senior management executive. She can be reached at [email protected].