Generative artificial intelligence (AI) is becoming pervasive in nearly every industry, and genealogy is no exception. As genealogists, we wish to understand how to incorporate AI models, like ChatGPT, Claude, and Gemini, into our genealogy work. But, before we do, what can we learn from other industries that are already using it?
As a marketing professor, I keep current by reading industry trade journals and scholarly research on a variety of marketing topics. As a genealogist, I believe there are lessons to be learned from all the successes and failures brands have made while marketing their products using AI.
AI Mishaps and Triumphs in Marketing
So, what lessons did I learn? While I promise not to get too professorial, my observations can be grouped into four areas: protecting brand assets, abusing creative license, pitting man vs. machine, and reducing human workloads.
Protecting Brand Assets
One trend I noticed is that brands run afoul when AI is used to artificially enhance product benefits. For example, an AI-generated image of Lego’s Ninjago toy was placed on its website compromising the toy maker’s reputation for authenticity among its passionate followers.[1] Lego uses AI in the product development process much like it might use any other tool or insight to generate ideas, but using AI to depict its unique product in final form misjudged consumers’ expectations.
To avoid issues like this, L’Oréal and Dove have each vowed not to use AI generated images of humans in its advertising to preemptively protect the benefits associated with their respective brand assets, which are closely aligned with the appearance of hair and skin.[2] In their view, to use AI to depict healthy skin and hair is akin to image manipulation or deception. They believe they need to show how their products work on real people.
Abusing Creative License
Controversies around brands’ use of AI have mostly been with image generation as it is the most visible form of AI to consumers. Instacart used AI-generated images of food to accompany recipes used to promote its grocery delivery service. However, as potential customers soon discovered, one of the images showed two cooked chickens conjoined at the shoulder and a hotdog possessing the texture of a tomato.[3] Fashion brand Selkie used an image of a dog with too many toes.[4] Contrary to popular belief, people do notice the tiny, albeit awkward, details.
Pitting Man vs. Machine
But not all missteps are directly related to faulty images. Some brands have been criticized as lazy or disingenuous for its use of AI in marketing. For example, Levi used AI to create more diverse models for its website,[5] and Vanderbilt University used AI to write an inclusivity statement for its Office of Equity, Diversity, and Inclusion.[6] Both laudable tasks with the best of intentions, but there are some tasks consumers believe humans should do not machines.
Reducing Human Workloads
Some great uses of AI in marketing are related to process improvement, product development, and data analysis. IBM used AI to standardize inconsistent customer service documentation.[7] Shake Shack used AI to search subreddits on the social platform Reddit for contextual conversations about chicken sandwiches and then used the data to target advertisements within the most relevant subreddits with great success.[8]
Despite some of the wins with data analysis, brands are concerned about data security and some have discovered inherent biases with the training data used to program their AI models. After a lawsuit made the issue more apparent, iTutor discovered its AI-powered recruiting software automatically rejected female applicants over the age of 55 and males over the age of 60, an artifact of the data used to train its model.[9]
Fears about copyright and data privacy issues led some brands to incorporate restrictive language into the contracts with their advertising agencies who use AI during the creative process.[10] Brands are leery about how their proprietary data and consumer insights are used to train the agency’s AI model, which might be used in the future with the brand’s competitors.
Applying AI Marketing Lessons to Genealogy
When using AI, genealogists should never compromise their value proposition or what makes them unique to potential clients or those they would seek to influence through their thought leadership. AI should be used to compliment or reinforce their core competencies or what they do well (skills, experience, and specialized knowledge).
Brands stumbled with their use of AI when they failed to protect and perhaps even underestimated what consumers valued in their product offerings. In the advertising projects my students create and with the small business owners I encounter, I often find that the excitement in one’s business often has them jumping straight into the marketing tactics without fully understanding their strategy. They fail to articulate their value proposition and understand their core competencies before engaging in customer acquisition activities.
I get it. Defining one’s business strategy is not as sexy as the customer-facing communication elements such as advertisements and social media posts. But when we write blog posts, create podcasts, and put our persona out there without a good understanding of how we are different from other genealogists (value proposition) and what skills set us apart (core competencies), our messages appear disjointed and likely fail to resonate with our intended targets. The use of AI in genealogy is no different.
Your Personal Genealogist Brand
For most genealogists, AI is probably not part of your value proposition – save the few genealogists who are leading the industry on understanding the latest news, trends, and applications affecting genealogy. AI is perhaps better suited to support our core competencies as a tool and elevate our value proposition, which is what Coca-Cola has recently done. Coke used AI to complement employee efforts to develop new flavors, create immersive consumer experiences, and generate advertising content reinforcing the brand’s emphasis on “optimism and happiness”, which is its value proposition.[11]
As such, it’s important for genealogists to protect their value proposition and reinforce their core competencies. Don’t have AI do what you should probably be doing yourself and what others value about you. There’s a fine line between using AI as a tool and AI replacing your ability to analyze and correlate data, resolve conflicting evidence, and write a sound conclusion, which are some of the foundations of the Genealogical Proof Standard.[12] Indeed the Board for Certification of Genealogists just announced their Interpretation Regarding the Use of AI in their application portfolios, which is more consistent with the view of AI as a tool rather than a substitution for human reasoning.
AI should be viewed a resource or skill to improve your abilities to perform research and draw conclusions. It is a tool inasmuch as using autosomal DNA, What Are the Odds, Newspapers.com, or any other specialized knowledge in a locale, record group, or archive helps us do our work. AI can help transcribe documents, analyze an image, decipher handwriting, create images, and interrogate data. However, it shouldn’t replace your ability to reason, correlate, and conclude. Protect your product assets – you.
AI still needs human intervention and probably always will. It needs to be trained, and its output reviewed. AI can reduce the time and energy associated with some routine, daily genealogy tasks, but only after careful vetting and practice.
AI Prompts
One of the challenging aspects of using AI are the prompts used to tell it what to do. Referred to as prompt engineering, it takes considerable effort to know how to make the command to receive the output you need.
Genealogy is about people, and if you need to make images with people, creating the right composition of gender, age, and ethnicity in the appropriate century is a challenge. Engineering the right prompt is an exercise in patience.
One piece of marketing collateral I discovered can help with prompt engineering when creating people are involved. It comes from Dove’s Real Beauty Prompt Playbook, which it recently authored “to create a more equitable, inclusive, and diverse representation of beauty and appearance”.[13] Dove understands that its value proposition is about developing products that support a woman’s beauty and redefining the beauty standards to be more inclusive. So, Dove developed the 72-page playbook to not only understand how prompts work but also to provide an “Inclusive Prompting Glossary” with examples by age, body type, hair, face, age, race/ethnicity, and more. If you’re creating genealogy images, whether of people or not, it’s a must-read.
Conclusion
There’s a lot of buzz around AI in both genealogy and elsewhere. This article was meant to get us thinking about the responsible use of AI especially as it pertains to our personal brand as genealogists. My commentary is not meant to be prescriptive or even a proclamation for or against the use of AI in genealogy. In full disclosure, I use AI to help generate some of the images on my website and the images used in some of my blog post headers, including the image in the header of this post.
It’s probably worthwhile to point out that the observations discussed in this article are likely more relevant for those genealogists where there is a value exchange between people – money or influence, like seeking client work or providing thought leadership through blogs, podcasts, social media posts, or other publications. For genealogists performing research for their own private consumption, they may not need a value proposition, but they still need to use AI responsibly and ethically.
To this last point, I believe it’s important for everyone to understand where our constituents in the genealogy space are with respect to the use of AI in genealogy and in general. There are movements to label creative efforts by people as “made by humans” or to attach a “Not By AI” badge to their work.[14] Whether for your clients, your blog readers, your social media followers, or your own family, we need to gauge what these groups view as acceptable uses for AI. While this is likely an evolving target, understanding how all applicable audiences view AI can better help genealogists articulate their value proposition and decide how or if AI should become part of their core competencies.
My last comment here is somewhat outside the purpose of this article and perhaps a better question for a future blog post. Nevertheless, as genealogists, we understand the need to cite the use of AI in our work, but why not pronounce our work or portions of it as “human-made” when it is in fact AI-free? Consumers already can’t tell whether content is AI-generated or not, and as AI becomes more pervasive, why not flip the paradigm and cite our work as, well, our work? Our byline may not be enough. We may need to be more specific and indicate which sections of our proof arguments and research reports were reasoned, correlated, and written by us. There’s value there.
Never miss a blog post. Subscribe now.
Sources
[1] Hiken, Asa (2024, April 19). “5 Brands Restricting AI in Their Marketing,” Advertising Age.
[2] Ibid, A. Hiken.
[3] Feger, Angie (2024, April 19), “ Lessons from Brand Missteps in Responsible AI Content Marketing,” EMARKETER, accessed 18 July 2024 at https://www.emarketer.com/content/3-lessons-brand-missteps-responsible-ai-content-marketing?utm_source=Newsletter&utm_medium=Email&utm_campaign=edaily+6.28.2024&utm_id=edaily+6.28.2024&utm_content=edaily+6.28.2024.
[4] Ibid, A. Feger.
[5] Ibid, A. Hiken.
[6] Pasquarelli, Adrianne (2023, December 13), “Biggest Brand Fails of 2023,” Advertising Age.
[7] Lebow, Sarah (2024, June 27), “How IBM, Rillet Turn Generative AI from Productivity Tool into Marketing Machine,” EMARKETER, accessed 17 July 2024 at https://www.emarketer.com/content/ibm-rillet-generative-ai-productivity-tool-marketing-machine?utm_source=Newsletter&utm_medium=Email&utm_campaign=edaily+6.28.2024&utm_id=edaily+6.28.2024&utm_content=edaily+6.28.2024.
[8] Ostwal, Trisha (2024, June 27), “Exclusive: How Shake Shack Used Reddit and AI to Drive Sales,” AdWeek, accessed 16 July 2024 at https://www.adweek.com/media/exclusive-how-shake-shack-used-reddit-and-ai-to-drive-sales/.
[9] Olavsrud, Thor (17 April 2024), “10 Famous AI Disasters,” CIO, accessed 16 July 2024 at https://www.cio.com/article/190888/5-famous-analytics-and-ai-disasters.html.
[10] Ibid, A. Hiken.
[11] Springer, Jon (2024 June 5), “5 Ways Coke is Using AI to Enhance Marketing,” Advertising Age. And Sloane, Garett (30 November 2023), “Top 5 AI Marketing Activations Brands Need to Know Right Now,” Advertising Age.
[12] Board for Certification of Genealogists (2023). Ethics and Standards. Accessed 18 July 2024 at https://bcgcertification.org/ethics-standards/.
[13] Unilever (2024). Beauty in the AI Age. Dove.com. Accessed 16 July 2024 at https://www.dove.com/us/en/stories/campaigns/keep-beauty-real.html.
[14] See for example, https://notbyai.fyi/ and Eaton, Meredith (2023 November 13), “Made by Humans: the New Gold Standard for Marketing and PR Output?” PRNEWS, accessed 16 July 2024 at https://www.prnewsonline.com/made-by-humans-the-new-gold-standard-for-marketing-and-pr-output/.
Acknowledgment: The image used within the header at the top of the blog post was created using Microsoft’s Copilot AI-powered assistant (DALL-E 3) and added to the title slide. AI tools were not used to generate the blog’s intellectual content or provide writing assistance. The post was authored solely by me.
Great post
Thanks! I know you’re doing a lot of great work in this area. Thanks for all you do — and share!