Navigating AI use Without all the Information
When topics of automation were discussed in the past, it was not uncommon to be under the impression that it would first and foremost target ‘low skill’ and undesirable jobs (think self-driving cars and trucking), while never being able to rival professionals in highly technical fields. Furthermore, even if we could automate every job in the world, we could all be happy as artists or musicians, as it would be inconceivable for a machine to produce something so intertwined with the human experience. Interestingly enough, in the past couple years with the so called ‘AI revolution’, LLM’s and other generative models powered by machine learning have flipped the script and now threaten the work of those once thought untouchable; leaving a bad taste in the mouths of not only those directly affected, but also a handful of consumers. Because this technology, no matter how big a splash it made, is still in its relative infancy, I thought it important to outline a few examples in which AI could potentially harm a brands image or performance.
Pitchforks and Paintbrushes
While the works of starving artists may have been invaluable to the western canon, many today prefer to operate on a full stomach, selling their intellectual property to companies either by commission or on their payroll as graphic designers. Since the development of generative AI images with models like ‘DALLE’, this technology has seemingly made its way into almost all of these consumer-built models, including ones such as ‘Chat-GPT’. This increase in availability came with it an increase in awareness; companies, big and small started reflecting on their bottom line after noticing that a machine could generating high quality images in mere moments and completely for free.
Just last year Coca-Cola released a series of advertisements for the holiday season, something integral to their brand image for decades, made entirely with AI. These ads felt like something of a test run for the company, as some limitations with the tech were put on display, such as the inability to create convincing facial expressions and maintain consistency beyond a few seconds. Technological concerns aside, this decision was met with some heavy backlash on social media, not only because it took away potential work for actors and set designers, but also because of just how impersonal and disingenuous the whole thing felt. Megan Cruz of The Broad Perspective Pod wrote: “This is always what AI was going to be used for btw. It’s not some great equalizer. It’s a way for already massively wealthy execs to add a few more mil to their annual bonuses by cutting creative teams entirely & having a machine vomit up the most boring slop imaginable instead.” If a titan of industry such as Coca-Cola received this level of backlash, imagine what could happen to any number of small businesses with this kind of press.
Innovation or Infringement?
While the court of public opinion may shift, federal courts are not known for their leniency. Because of just how new AI is, the landscape is reminiscent of wild west. Federal restrictions are almost nonexistent; however, the business software alliance estimates that over 700 bills have been passed on the state level to reel in the industry. With the severity states are handling this, it is only a matter of time until national regulations begin forming, and generative content is something being examined particularly closely.
To understand why lawmakers have their sights set on generated content, it is first necessary to have a general understanding of how these algorithms create the content they do. rather than genuine intelligence, machine learning involves ‘training’ an algorithm by feeding it incomprehensible amounts of information as it identifies patterns uses them to imitate the best possible response or solution. Chat-GPT has never ‘seen’ a van Gogh but from an understanding of his brush strokes and use of bright contrasting colors is able to create convincing recreations. Calamity ensues when unlicensed art makes its way into AI Training data, potentially violating copyright laws and calling into question the ownership of these AI tools creations. Some of these issues are being investigated at this very moment, particularly in the landmark case “Andersen v. Stability AI et al”, but countless more pop up seemingly daily. While it is unlikely someone would be knocking on your door just for some fluff images, everything is subject to change, and it is smart to keep your finger on the pulse of the industry as everything unfolds.
Robots Tanking Rankings
LLM’s work almost identically to the machine learning we’ve discussed in the previous section, only they do so with text as opposed to images. This has had a colossal impact on search engine optimization (SEO). This is an incredibly nuanced field that has a whole host of moving parts but at its most basic level: When google looks to show you a website based off your search it does so based on not only your own habits, but how convincingly websites are able to engage consumers and the likelihood they have what you’re looking for. A large way in how they achieve this is via keywords (say if I search banana, Chiquita will pop up because of how many times the words ‘banana’ shows up in their site). Companies have been looking for ways to abuse this system for years, and right now it’s easier than ever; if I asked Chat-GPT to generate text with care to include the keywords ‘x, y, z’, then I can boost the chances of my site popping up tremendously.
This, like so many topics relating to this industry, is a touchy subject at the moment. Because of how efficient this is, websites optimized with keywords using AI are ranking higher than ever, so much so that people are starting to take notice. Just last year, WIRED published an article discussing how “Google Search Ranks AI Spam Above Original Reporting in News Results”. In the short term this keyword stuffing can lead to a lack of SEO value because the AI generated content and end up being too similar to another site using the same strategy, oversaturating the search page and ultimately leading to sites being bogged down as ‘perfectly optimized’ becomes the standard. While in the long term, this could even be a detriment to a website as SEO itself changes. Google have been clear about their intentions to rank content based on E-E-A-T (expertise, experience, authoritativeness, and trustworthiness). And future updates to the software could cause a genuine website using keyword stuffing to rank lower.
While it can be so easy to focus on the negatives, AI has done a lot of good and can be an invaluable tool, especially for small business who lack the manpower of the big guys. It is best, however, to understand the growing pains of AI and what they can impact, learn to use the technology responsibly, and to understand it is no substitute for human creativity or ingenuity.
Recent Comments