(Originally Posted 01/2023)
Sometimes it’s hard to tell if the news and information that comes across my feed is relevant or not. The proliferation of technology in our daily lives has made it increasingly difficult to discern what information is relevant and important, and what is simply noise. The constant stream of posts on social media, combined with the rise of "influencers" using pithy copy and hashtags to promote their content, has created an "Ouroboros" feedback loop where the information we consume ends up consuming us.
In most cases I prefer to maintain a moderate level of skepticism. I hate to think I live in an echo chamber of one. However, it wasn’t until Christmas day when I opened the Sunday Business section of The New York Times did I realize that what was once my own novel interest had expanded into the public’s view. When I saw the topic again pop up in the Financial Times a week and a half later, there was no denying. It was in fact “trending”.
This trending topic is ChatGPT. Developed by the San Francisco research institute OpenAI, ChatGPT is a language generation model. This “Generative Pre-training Transformer” is a type of Artificial Intelligence machine learning model and neural network that generates natural language responses, indistinguishable from text written by humans, similar in style and content to a given input or query. The “pre-training” part means that the model is trained on large datasets, consisting of billions of words from articles, books, and websites, before being tuned for a specific purpose. Right now, this beta version of ChatGPT is in its third iteration. Its most recent update was on the 9th of January. It’s worth noting that in 2019 Microsoft invested $1 Billion into OpenAI, and recent reports from Reuters and Bloomberg suggest an additional $10 Billion investment.
So what can ChatGPT do? Let’s say you want to know the foundational differences between Hinduism and Buddhism for a school paper, or how to structure your 5-day fitness workout schedule for your 2023 New Years resolution. Or maybe you want to know how to write Adobe Flash code for a multi-photo carousel to use on your blog but you aren't a programmer. Simple. What about something a bit more abstract? Ok. How about creating a Warren Buffet inspired investment strategy for the S&P 500 with stock suggestions based on their performance over the last 20 years. Still not abstract enough? Maybe you’re a big fan of Canadian clinical psychologist and social critic Dr. Jordan B. Peterson, or you are the good doctor, so you ask it to create a follow-up 13th rule to your NYT best seller “Beyond Order, 12 More Rules for Life”, written in the style that combines the King James Bible and the Tao Te Ching. Its’ response, given almost instantaneously, could look something like this,
“Get wisdom, get understanding: forget it not; neither decline from the words of my mouth. Forsake her not, and she shall preserve thee: love her, and she shall keep thee. Wisdom is the principal thing; therefore get wisdom: and with all thy getting get understanding."
"The wise man does not lay up his own treasures. The more he gives to others, the more he has for his own." - Tao Te Ching, Chapter 81”
You can cross reference the output with your own King James Bible if you’d like, but I’ll just tell you here - it nailed it (no pun intended). Pretty sophisticated stuff. Again, this is the beta version. Don’t act like you’re not impressed.
After three weeks of playing with ChatGPT-3 I’ve only scraped the surface of its capabilities. Its immediate applicability is hard to ignore. I used it to create a 38 page “Strategic Organizational Framework” document for a local non-profit. Something that would have normally taken weeks to prepare was completed in 36 hours. The value of AI with its combination of machine learning and large neural networks is both immediate and immense. And therein lies the rub!
Obviously, I’m not the first person to have a healthy fear of AI. Growing up throughout three decades of action/sci-fi movies, encompassing the likes of “The Terminator '' and “The Matrix” trilogies, as well as the classics like 2001: A Space Odyssey, I’ve heard this siren call before. We are on the cusp of AI tech infiltrating many aspects of our life.
“It's a familiar tune that's hard to ignore, the promise of powerful technology that could change the world. But like the sailors of old, I know that the alluring song of the siren can also lead to destruction and disaster if we're not careful” - ChatGPT
The use of AI isn’t exactly breaking news. The process of using computers to help create media has been around since the 1970s. In 2016, a team of researchers from J. Walter Thompson Amsterdam, Dutch bank ING, and The Mauritshuis museum created a machine learning model that was trained on a dataset of Rembrandt's paintings. The resulting painting, called "The Next Rembrandt," was unveiled to widespread media attention.
More recently, programmatic ad exchanges used by many large media companies like Meta, Publicis Group, and Hearst Corporation, all use computer modeling and AI tech for their ad buying strategies. These machine learning algorithms can take into account a wide range of data, such as demographics, browsing history, purchase history, and even social media activity, to identify patterns and insights that can be used to optimize ad targeting and improve ad performance.
In the United States, revenues from artificial intelligence software are expected to double from just under $25 Billion in 2022 to over $50 billion by 2025. A 2020 survey based on 303 marketers showed that 32% already used AI to create ads. Of those ads, 60% were used for digital banners and 54% were for Social Media posts.
AI is being integrated into various aspects of its business model, including copywriting (Jasper AI. Inc.), workflow automation (Zapier Inc.), image creation (Dall-E, Midjourney), text-to-speech voice replication (Vall-E), and the use of computer-generated fully animated human avatars (Synthesia).
While much of the discussion on the use of AI in advertising has been oriented around job security. There is real concern that AI-driven advertising could lead to massive job losses as the technology advances, however, my concerns are more centered around the ethical and moral implications of this technology’s use. The ability to generate highly targeted and persuasive content at scale, AI can potentially be used to influence people's perceptions and beliefs on a massive level.
Firstly, as I already alluded to, AI-driven advertising may lead to increased surveillance and manipulation of consumers by tracking behavior, creating personalized ads, and target certain demographics with greater precision. This could lead to greater control over consumer behavior and decision-making, and ultimately manipulation of public opinion and a loss of consumer autonomy. In addition, AI-driven advertising may exacerbate existing inequalities. Many of these models still rely heavily on data input by humans who may pass on their inherent biases into these new technologies. Furthermore, AI-driven advertising could be concentrated to the larger, more well-funded organizations over smaller companies as the latter would have less capital to invest in technology and data. Lastly, these AI-data driven models could inadvertently and disproportionately benefit certain demographic groups while excluding others. Therefore it is important to consider the potential consequences of using AI for advertising media and to ensure that its use is guided by ethical principles.
I argue private companies, rather than governments, should be leading the way to establish the moral and ethical guardrails for AI in advertising. Private companies are typically more nimble than governments and can respond faster to changes in the market and technology. Similarly, the private sector, driven by competition and innovation, could help ensure ethical guidelines are constantly improving and adapting. They have an inherent interest in maintaining public trust and avoiding negative publicity. This creates an incentive for them to establish ethical guidelines that protect both consumers and their own reputation. All of this allows them to stay ahead of the curve before problems arise.
It's not to say that governments don’t play a role in the regulation of these technologies. However, governments have limited resources and other priorities, which can make it difficult for them to effectively and efficiently monitor or regulate the technology. Therefore it's best done in partnership with private companies. This collaboration between the private and public sectors could ensure that the use of AI in advertising respects human rights and promotes the common good. A good place to start could be with Microsoft’s own “Responsible AI Standard”.
Maybe we’re not a step closer to “Skynet” or waking up in an artificial womb covered in pink translucent goo, but many industries and organizations are already using these technologies with more on the way. The development of advanced AI technology, like ChatGPT, has the potential to revolutionize many industries. But none is more powerful than the media industry which can shape and reshape national culture and consumer behavior. Ultimately, it's the responsibility of the private sector to ensure the benefits of AI in advertising are achieved without harming or undermining consumer autonomy or privacy.