Here’s What Worries Me About AI

Posted on by Larry

The AI genie is fully out of the bottle. AI is a significant, watershed event for the creative arts. AI is powerful, expandable, teachable and applicable to multiple industries. Hundreds of new AI applications are released each week. The only things we can do now to control AI is to regulate it and/or accurately identify it.

Regulation is possible, but it won’t be quick. Finding a way to ID every AI-created work will be impossible. Someone will always find or need a workaround.

The copyright office ruled a couple of weeks ago that AI generated media can not be copyrighted because it was not created by humans. (National Law Review)

PCMag reported that “multiple employees of Samsung’s Korea-based semiconductor business plugged lines of confidential code into ChatGPT, effectively leaking corporate secrets that could be included in the chatbot’s future responses to other people around the world.” Anything entered into ChatGPT (but not the API) is added to the training model available to all users, thus making it public. (PCMag.com)

In the past, automation destroyed tens of thousands of blue-collar jobs. Now, the shoe is on the other foot, white-collar jobs are threatened. “Generative AI tools could completely change the way people find and synthesize information, replace or disrupt hundreds of millions of jobs and further cement the power Big Tech companies wield over society.

“‘People talk about AI as a technological revolution. It’s even bigger than that,’ said Sam Altman, founder of OpenAI. ‘It’s going to be this whole thing that touches all aspects of society.’” (Washington Post)

Today, only a few months after the exploding launch of generative AI, we see the following trends:

But, it will devalue creative work because how valuable is an image, or video, that can be created in scant minutes from a paragraph of text?

It will also cost jobs; especially jobs in the creative industries. This is the part that troubles me. Many tasks in media don’t require great amounts of creativity – but they do require specific skills to accomplish a task. When those skills are automated, many of us become unnecessary.

NOTE: I should also mention, though this has less impact on image and video, that generative AI can deliver believable results filled with gibberish, or make assertions that sound true but are false, even criminally wrong. In other words, its results can’t be trusted.

(Photo courtesy of Tara Winstead – Pexels.com)

I recently had a conversation with the CEO of a visual effects company. He’s excited about applying AI technology to his products because it enables them to do things that they could never do before. It’s the “teachability” of AI that excites him.

A New York Times headline read: “Instant Videos Could Represent the Next Leap in A.I. Technology.” Here’s the key takeaway:

“Runway… is one of several companies building artificial intelligence technology that will soon let people generate videos simply by typing several words into a box on a computer screen.

“The ability to edit and manipulate film and video is nothing new, of course. Filmmakers have been doing it for more than a century…. But systems like the one Runway has created could, in time, replace editing skills with the press of a button.”

And THAT is my concern.

Philip Hodgetts commented that Synthesia.io, which makes synthetic animated human faces to act as on-camera presenters, now claims 6,000 customers and “40,000 teams” using their synthetic avatars for talking heads.

At a minimum that’s 6,000 lost jobs for presenters, camera operators, sound and makeup people because typing a few words into a web browser is easier than setting up camera, audio, lights, and hoping the neighbors don’t make noise! $30 a month is a compelling alternative for creating 10 minutes of production.

AI-enabled software means that developers of creative tools no longer need to sell solely into the limited professional market. Instead they can sell into the vastly larger consumer and corporate markets because no professional media skills are needed to use what used to be technically complex software.

Today artists use the content-aware tools in Photoshop to seamlessly remove objects from an image, or select a subject to remove it from the background. Last week, Meta demoed a tool that selects any object out of any background without requiring Photoshop.

A tool that anyone can use.

The bulk of our industry, where many of us are employed, is not the Hollywood (or Bollywood) block-buster or high-end commercial production. It’s corporate, educational, marketing and personal communication in all its forms. The high-end will always exist, but that’s not where most of us work.

When an executive assistant, with no media training, can type in a prompt and get a photo-realistic image back in seconds, why would that company hire an artist to do it at far more cost taking far more time? Creating effective prompts may require practice and skill, but it does not require a media background.

When a marketing prompt expert can create an ad by typing a text prompt which creates a video with people who move and talk and look real, where does that put the companies that specialize in advertising or the actors in them?   Sure, the results today are a bit stilted, but every rejection teaches the AI system how to do it better. Its greatest strength is that it can “learn.”

When you can creates videos by typing text, rather than the labor intensive process of video production, how many clients will insist on spending more money and time to do it “the old fashioned way?”

High-end movie production – especially visual effects – is likely to be profoundly altered by these new automated tools. You don’t need legions of rotoscope artists when AI-assisted software can extract subjects and replace backgrounds in seconds; no skill required.

Granted, we aren’t at this level – yet. But given the speed of change, all of this should easily be possible in a year… or two.

Even musical performances can be quickly created using a new soundtrack, existing footage and AI-technology. Lip-sync has always been used in creating video.

Thinking about the near-term, the video production areas that seem “safest” are weddings and school plays (because the people in them are uniquely identifiable) and live events. Everything else is likely to see significant disruption from AI.

My fear is that, like blacksmiths when the automobile became popular, we will have skills that no one needs because technology has moved past us. Maybe I’m wrong. But what if I’m right? How do we plan for the future, when the foundation upon which we built our careers is shifting so quickly?

AI is great for developers because it vastly expands their customer base and allows them to do what they were not able to do before. End users love AI because it completely redefines the skills needed to create high-quality visuals. You no longer need to be a media professional to create professional results.

This is a sobering thought. Yes, AI is an exciting new frontier, but this rising tide does not lift all boats. There will always be a need for blacksmiths, just not as many of them. That’s what worries me.

EXTRA CREDIT

Here’s a “Truthfulness Challenge” that media producers need to face in the very near future.


Bookmark the permalink.

19 Responses to Here’s What Worries Me About AI

  1. This is so reminiscent of the birth of PowerPoint. When “speaker support” graphic design landed on administrative desks it foretold the rise of truly mediocre graphics. It’s what desktop publishing did to the art of designing with typography. This time it feels like the professional community will need to embrace certain guardrails to preserve the qualities and ethical standards of our work.

    • Larry says:

      Jonathan:

      Reminiscent, but with a much broader impact.

      Your desktop publishing analogy is apt, though. Once desktop publishing – creating text and graphics on a computer and outputting to a laser printer – got established, small letterpress and print shops were replaced by Kinko’s with Xerox copiers. The printing industry never recovered.

      Larry

  2. Jon May says:

    The younger generation is too quick to embrace the “cool factor”over practicality and the looming, and very real downside, for which once the genie is out of the bottle, they aren’t going back. We once, not too long ago, had personal privacy. Now, with tools available to anyone, with someone’s name, I can find out their age, where they live, their family and relatives, email, phone numbers, I can get a picture of your home, and if it’s ever been on the market, interior pictures of every room of your home. Add to that, Facebook and Google data mine every aspect of your life, and YOU have no privacy anymore. Not a shred. The developers of these technologies never considered the proprietary issues of personal privacy, instead embracing the “cool factor.” Now, you have nothing left. Sadly, we cannot trust that AI developers will include responsibility as part of a trusted technology. They already gave away your personal privacy, what else do you have to lose? Already, computers are making hiring and firing decisions; your entire career path is decided by microchips. We have surrendered and abandoned practicality, personal resourcefulness, decision making and the propriety of technology, just because some developer finds it “cool.” Zager and Evans nailed it. (millennials may have to look it up.)

  3. Bob Aiese says:

    It feels like we are all walking into a mine field of law suits and litigations. If no one owns the rights to an image does everything AI generated become public domain?

    • Larry says:

      Bob:

      Yes. That’s what the engineers at Samsung discovered. They posted proprietary code to ChatGPT, which incorporated it into its learning engine, available for use by other companies.

      Additionally, since AI-generated work can’t be copyrighted, by definition, all AI work is public domain. Naturally, this makes AI open season for lawsuits, especially regarding images used for training an AI system. Getty has already sued OpenAI about this. More will follow.

      Larry

    • H. Nelson says:

      Just as an aside, the “new rules” of the Copyright Office are not yet law but, rather, laid out in a 12 page “policy” that gives guidelines on registering work that ranges from 100% machine generated, to works wherein “a human may select or arrange AI-generated material in a sufficiently creative way that ‘the
      resulting work as a whole constitutes an original work of authorship.’ Or an artist may modify material originally generated by AI technology to such a degree that the modifications meet the standard for copyright protection.” They made it clear that much of this will have to be handled in a case-by-case fashion.

      • Larry says:

        H. Nelson:

        Thanks for the clarification. I should have made clear this was not yet law. The key point, though, is that at the moment AI-generated material is public domain; as the engineers from Samsung discovered to their distress.

        I have ZERO doubt that all of this will spend a long time in lawsuits and policy discussions. For now, though, folks who make a living creating artistic work need to pay special attention to how they use and sell AI-generated art.

        Larry

        • H. Nelson says:

          Absolutely. The policy also stated that it’s up to the applicant to reveal that the content is machine generated or machine generated assisted. And knowing human nature, I’m sure that is going to be taken advantage (or abused) and lead to plenty of court battles!

  4. H. Nelson says:

    “High-end movie production – especially visual effects – is likely to be profoundly altered by these new automated tools. You don’t need legions of rotoscope artists when AI-assisted software can extract subjects and replace backgrounds in seconds; no skill required.”

    I worked in visual effects (VFX)from the late 80s to the early 2000’s and saw, firsthand, the transition from photochemical/optical VFX to CGI. It was brutal for many people like model-makers, puppeteers and traditional rotoscope artists. Many had to learn new skills or get left behind. I saw the same thing happen with friends who worked as matte painters & traditional 2D animators. And yes, some jobs were eliminated, some changed the tools they used & new jobs were created as well.

    The end result was, of course, a mixed bag. Giant motion picture labs like Technicolor & Deluxe either closed or moved into different arenas of post-production. The small mom & pop lab Fotokem rose to be the only real player left in that field today.

    The original promise of CGI was going to streamline VFX production! & be more cost effective! Now we have armies of digital artists (many doing grunt-like repetitive tasks) and VFX budgets have exploded. I’m sure AI will help to rein in the costs in VFX and people who can write effective prompts (or whatever the input evolves into) and THAT will create new jobs. It’s a never-ending cycle for sure. But there are jobs in VFX that really are just a grind for a paycheck. You get toward the release date of a film and your pushing people 10-15 hours a day, and sometimes working 7 days a week. It ain’t good & if you could find an AI solution instead of working people half to death, it’s tempting to say “Yes!”

    The big difference now with AI is how much of the work are we going to allow the machines to do? A lot of mediocre content gets generated when people without skills or artistic ability are removed from the creative process. Hollywood jumped on the CGI bandwagon and there is a LOT of bad/cheap VFX that audiences did not respond to. It’s up to us to keep our skills sharp & pay attention–we don’t know which way this trend is going to go, but the beginning will be a rough road, and we have to be ready for anything.

    • Larry says:

      H. Nelson:

      Thank you for an excellent comment and a perspective on the VFX industry that I don’t have.

      Larry

      • H. Nelson says:

        Thank YOU, Larry! These industry-wide transitions are not easy & incite a lot of FEAR. When CGI/Digital reared it’s head, many talented people in the VFX industry had not only never used a computer before, but had to figure out how to use these rather crude still-evolving tools to match what they could better & faster to by hand. Some adapted, others gave up & moved on.

        Someone like yourself, who is in the trenches, keeps up on the latest technology & trends and also has the ability to explain things in a very clear, pragmatic & down-to-earth way can help alleviate a lot the fear of the on-coming change. Thanks for all you do for our industry! 🙂

  5. Gregg Hall says:

    Even if we shift our focus to live events, AI is already eroding opportunities there as well. Take for example traceup.com. This technology is designed to capture amateur sports with AI powered cameras and editing combined. We can look at the footage this generates today and see the shortcomings. However, it WILL get better over time, and as it does it will be embraced in corporate settings as well. It’s difficult to see how we wiggle out of this predicament.

    • Larry says:

      Gregg:

      Interesting – I had not heard of TraceUp.com. I will check them out.

      And, yes, “It’s difficult to see how we wiggle out of this predicament.” is an understatement.

      Larry

  6. Emerson says:

    Hi Larry I think is a bit naive that people think that their lives are going to be the same for decades especially if all human knowledge doubles every 10 years and everything moves exponentially…and I don’t believe in the “prophets of doom”…We will adapt, we always do…I think AI is a tool for us to use…using a template is not kind of using somebody’s work just to save time for us to do our videos quickly? maybe in the future, we will all be “AI prompters”…who knows? we just need to embrace change, learn and adapt…

  7. Michael Powles says:

    AI for voices is a very dangerous thing! As you illustrate in your article, Larry: “6,000 lost jobs for presenters”. My voice is totally unique to me and my intellectual property.

    But I know of an example in the US where a voice-over artist (VO) recorded work for a client who promptly hijacked the VO’s work to create the client’s own AI VO material for his company so they didn’t have to employ the human VO anymore. Is THAT fair? Perhaps ‘fairness’ doesn’t come into it? This is what, in the UK, we’d call ‘sharp practice’ The rights of the human VO have to be protected by law.

    And what happens when making a phone call to one’s bank where: “Your voice is your password”! But is it YOUR voice that the computer hears!!!

    Larry adds: Michael is a professional voice-over artist, based in the UK.

    • Larry says:

      Michael:

      I agree – you should own the rights to your own voice and intellectual property unless you properly assign it somewhere else. Things get far murkier when your voice – or any other creative act – is used to train an AI system. You don’t create the finished work, but without your voice to train it, the finished results would not be the same.

      I think the courts will be very, very busy with this for several years. And crafting meaningful legislation will be hard.

      Larry

  8. Carlos Ziade says:

    This reminds me when we had a dark room, then later a Fujifilm photo printing lab in our photo studio back in the late 1990s, and that has gradually started to become unnecessary with the emergence of digital cameras and later, social media platforms, until we totally shut down our studio around 2010.

Leave a Reply

Your email address will not be published. Required fields are marked *

Larry Recommends:

FCPX Complete

NEW & Updated!

Edit smarter with Larry’s latest training, all available in our store.

Access over 1,900 on-demand video editing courses. Become a member of our Video Training Library today!

JOIN NOW

Subscribe to Larry's FREE weekly newsletter and save 10%
on your first purchase.