Features March/April 2024

Contracting Out and Partnering with AI

By Daniel L. Hocutt | Member and Ann Hill Duin | Member

What’s a technical communicator to do in response to generative AI? We recommend a couple of directions.

In what Ethan Mollick (2024) calls this AI Moment, we sample ways that technical communicators are contracting out and partnering with AI, blurring the lines between human and machine activity. In response to blurring these lines, we offer responses to the question “What is the technical communicator to do?”

What is being contracted out to AI?

Daniel works as a web content manager on the marketing and engagement team of the continuing education unit of a private liberal arts university. He approaches his position as a technical communicator. In this role, he manages the school’s paid advertising efforts on search engines and social media platforms. He’s a member of the LinkedIn Advisor network and regularly receives requests for expertise on user experience surveys.

A recent user experience survey asked questions about a new generative artificial intelligence (AI) tool that LinkedIn advertising rolled out. The tool, like those provided on other search and social advertising platforms, follows a similar chat pattern to AI tools like ChatGPT and Bard by soliciting a prompt and generating a response to the prompt. The prompt requested is a landing page URL for the product or service being advertised. The response, without requesting additional information, proposes a complete ad campaign for the product or service, including recommended conversion measurement (i.e., call to action, like a purchase, sign-up, or download), optimal budget, target audience, and a set of ad assets (headlines, descriptions, images, etc.). While the advertiser is responsible for vetting and approving each element of the advertising campaign on the ad platform, generative AI produces the bulk of the advertising content.

The intellectual, rhetorical labor of creating a digital advertising campaign traditionally falls to a marketing communication specialist who uses subject matter expert (SME)-provided tools like a detailed product description, results of target audience analysis, and a key competitors matrix to request or generate image sets, target audience interest profiles, ad copy, and advertising budget. Generative AI tools now in various stages of development and use on LinkedIn (https://www.linkedin.com/pulse/linkedin-introduces-ai-powered-ad-creation-tool-jonathan-pollinger-), Google (https://ads.google.com/intl/en_us/home/campaigns/ai-powered-ad-solutions), and Meta (https://www.facebook.com/business/news/generative-ai-features-for-ads-coming-to-all-advertisers) advertising platforms essentially contract out the often tedious and detail-oriented work of creating a digital advertising campaign within a recommended budget. Using existing machine learning (ML) tools, these advertising platforms are also able to optimize efforts in real time, adjusting keyword bids, spend levels, and audience profiles to maximize conversions within parameters recommended by generative AI and approved by the advertiser. While we’re aware there’s a circular logic at work—the platform optimizes using ML to its own AI-generated recommendations—the reality we’re demonstrating is that communicators can now contract out advertising’s generative composing processes to AI-augmented platforms.

What (who) are we partnering with in terms of AI?

AI is also birthing digital humans that may look and sound like us. The intellectual, rhetorical labor of designing the interface, again traditionally a part of the technical communicator’s portfolio, may now also involve designing and/or partnering with digital employees and ambassadors, as chatbots evolve into digital humans. In the Gartner organization’s discussion of 2021 hype cycle themes, Kasey Panetta wrote:

Consider digital humans, which are digital-twin representations of people. This technology presents an opportunity for licensed personas that enable new revenue streams. They can appear as avatars, humanoid robots or conversational user interfaces, like chatbots or smart speakers. These interactive, AI-driven representations seem human and behave in “humanlike” ways supported by a range of technologies including conversational UI, CGI and 3D real-time autonomous animation. (Panetta 2021)

Tracing the turn to human-autonomy teaming, last year Ann co-authored the book Augmentation Technologies and Artificial Intelligence in Technical Communication: Designing Ethical Futures with Isabel Pedersen, emphasizing that while “digital employees mimic human behaviors on the one hand, they also automate work tasks on the other with extensive AI capabilities that humans would not be able to perform… Through proxy agency, digital employees augment human cognitive capabilities” (Duin and Pedersen 2023, 194). These digital employees are designed to collaborate with us and are also sold by third-party companies for enterprise services. They emote verbal, facial, and gestural responses, with many being equipped to sense or detect emotion. As technical communicators, we need to understand their roles. “They do the work of humans; they collaborate with us; they learn from and with us” (195).

As an example of the use of ChatGPT in support of custom, interactive AI avatars, visit Soul Machines (https://www.soulmachines.com) and converse with any number of digital humans. “Its use cases span a broad spectrum that includes customer service, healthcare, financial services, brand ambassadors, entertainment, education, real estate, policing, and telecommunications/call centers. In the education sector, soul machines are positioned to answer financial, course, career, and guidance questions; to serve as administrative or virtual teaching assistants; and are ready to speak in 12 languages” (Duin and Pedersen 2023, 207). According to the April 2023 issue of the Harvard Business Review, “Digital humans are already making real money for their employers. Soul Machines… has upwards of 50 digital humans deployed in organizations around the world” (Seymour et al. 2023,. 11).

What lines are being blurred?

One of the major lines being blurred by the integration of generative AI is that between marketing communication and technical communication. In a special issue of Intercom on the future of technical communication, Saunders (2018) noted this blurring of lines in terms of content, while Earley (2018) identified the blurring of lines in terms of audience analysis. In the generative AI tools in use by digital advertising platforms, the technical writer’s work of content development and management, represented by the SEO-optimized structured content on a website landing page, becomes the prompt that generates the advertising campaign on the ad platform, complete with content and target audience analysis. From the careful work of technical communicators working with SMEs to develop and optimize web content, to the careful work of marketing communicators working with those same SMEs to develop target audience and competitive analyses, emerges the AI-generated advertising campaign, complete with recommended target, budget, imagery, ad copy, and placement.

In Daniel’s case, these roles are combined into a single content management professional. His technical communication responsibilities include information architecture, content writing and management, search engine optimization (SEO), UX design within the constraints of institutionally designed and managed templates, and development of marketing landing pages for educational programs. In many cases he collaborates with other team members and third-party providers to develop assets for brand and visual identity, to write content for use across digital and analog outlets, and to deploy technical solutions like website analytics and tags. His marketing communication responsibilities include digital advertising, target audience analysis, market and competitor analysis, and ad content writing. The lines between these roles regularly blur to obscurity. For example, the structured content authoring required for successful SEO provides the content and keywords required for search advertising and supports higher placement of ads on search engine results pages (SERP). The rhetorical, user-focused approach to audience and purpose in composing web content provides the audience analysis needed to specify target audience demographics and psychographics in social media advertising.

By integrating generative AI into digital advertising platforms, the roles of marketing communication and technical communication blur even further. Marketing communicators need technical communication’s approach to solving problems for users whose characteristics are being assumed by AI. Technical communicators need marketing communication’s approach to key performance metrics, goals, and measurement to ensure that users’ needs are being met when user data is the primary asset used to determine the effectiveness of AI-driven communication efforts.

Moreover, we contend that as technical communicators design and partner with digital humans via AI functional technologies, digital employees will provide customer service across multiple sectors as they are tireless; moreover, they are proposed to “bring meaningful connection to the digital world, where empathy and compassion have disappeared from customer interactions” (see UNeeQ, https://www.digitalhumans.com). As compared to previous technologies, digital employees bring the greatest autonomy and agency to date to the agent. Using tools such as those shared at Amelia (https://amelia.ai), Synthesia (https://www.synthesia.io/tools/digital-human), and Unreal Engine (https://www.unrealengine.com/en-US/license), technical communicators can create and customize digital employees using natural language. As we interact with digital employees, this expands our understanding and design of the interface between human and machine.

So, what is the technical communicator to do?

Given the current and emerging AI landscape, we offer the following suggestions for technical communicators facing this question: In light of generative AI’s integration into communication practices, what is the technical communicator to do?

Ensure human-in-the-loop (HITL) integration into emerging AI-centered workflows

At a recent higher education marketing conference (UPCEA MEMS 2023), Frederick T. Wehrle, Assistant Dean for Academic Innovation and Learning at UCLA Extension, recommended that higher education marketers spend 20% of their time learning about and experimenting with generative AI to identify its strengths and its pitfalls. We’d extend that recommendation to technical communicators as well. By understanding AI’s limitations, we’ll be in a position to advocate for and against specific uses of generative AI recommended or mandated by institutional policies. This is one way to practice HITL approaches to AI.

AI is built on the results of ML at massive scale. The large language models (LLM) used as training sets give AI its generative capabilities, but it’s important to recognize that machine learning is built on data inputs. LLMs are simply (incredibly) large data sets, and the adages about “good data in, good data out” and “garbage in, garbage out” remain accurate. Humans are involved in curating data for ML analysis. We may be assisted by algorithms and bots, but humans retain agency in determining whether data is adequately curated (which can be read as “ethical, representative, equally accessible”) for ML processing. One practical skill to meet this need to be involved in data curation is to learn basic data analytics practices to better understand the role data curation plays in ML and AI.

At the other end of generative AI integration are its results. Whether evaluating the recommendations from an advertising platform’s AI-generated advertising campaign or analyzing suggestions offered by an AI assistant, technical communicators can apply their usability and UX testing skills to sample and test the user experience of content generated by AI (see Hocutt, Ranade, and Verhulsdonck 2022). Technical communicators advocate for users. AI can’t serve as an advocate, at least not yet. Humans can question results, advocate for change, and retain the human user as the focal point of generative AI content. Applying usability testing methods to AI-generated content is another practical way to practice HITL. In short, technical communicators must be positioned to intervene throughout the design, adoption and adaptation of AI technologies. As Duin and Pedersen emphasize,

Audience is now an augmented audience; non-human agents now function as independent team members; and content is quickly and vastly produced through machine learning. As a result, TPCs [technical and professional communicators] must function as explainers, intervening to correctly distinguish correct information from misinformation based on overall understanding of the social and cultural context surrounding deployment of augmentation technologies. TPCs must also build understanding of emotional intelligence as it relates to the data used to train AI models. As TPCs work with AI algorithms, we can intervene to be better curators and stewards of the data used to train AI models as well as how that data then impacts those interacting with digital employees. For such intervention, TPCs need to build understanding of ethical algorithmic impact assessment tools and processes to help guide the design of and collaboration with augmentation technologies. (Duin and Pedersen 2023, 214)

As Elizabeth Spiers writes in a recent New York Times (January 7, 2024) opinion piece, “for now, it [AI, ChatGPT] needs adult supervision.”

Expand the blurring of lines among all who communicate and design using AI

We argue that blurring lines among communication professionals who engage with generative AI is a positive step in our response to AI. This article has identified marketing communication and technical communication as two fields where those lines are readily blurred. Other fields are emerging. Consider the way AI is being integrated into learning management systems (see https://www.proprofstraining.com/blog/ai-lms) as a natural language processing teaching and learning assistant. Educators will be questioning whether integrating generative AI into their course sites is a positive move. Marketing communicators and technical communicators can contribute methods for audience analysis and usability testing to help educators make informed decisions. Educators can provide methods for teaching students how to integrate AI into their workflows. The benefits of professionals in these fields collaborating and sharing their experiences centered around AI integration seem to far outweigh any doubts or concerns about disciplinary boundaries that get crossed or erased. AI blurs lines between professions and between humans and machines. Rather than resisting and retaining boundaries, we can center our efforts around AI to generate the best possible rendering and the best possible results for the prompts we provide and the people we assist.

Generative AI is here to stay. Technical communicators have an opportunity to partner with communicators and educators from across professional and disciplinary boundaries in order to serve as AI’s adult supervision and evolving partner.

References

Duin, Ann Hill, and Isabel Pedersen. Augmentation Technologies and Artificial Intelligence in Technical Communication: Designing Ethical Futures. Philadelphia: Routledge, 2023.

Earley, Seth. “AI, Chatbots, and Content, Oh My! (Or Technical Writers are Doomed—To Lifelong Employment).” Intercom 64, no. 1 (2018): 12–14. https://www.stc.org/intercom/download/2018-1/

Hocutt, Daniel. L, Nupoor Ranade, and Gustav Verhulsdonck. “Localizing Content: The Roles of Technical & Professional Communicators and Machine Learning in Personalized Chatbot Responses.” Technical Communication 69, no. 4 (2022): 114–131. https://doi.org/10.55177/tc148396

Mollick, Ethan. “Signs and Portents: Some Hints About What the Next Year of AI Looks Like [Blog].” One Useful Thing, (January 6, 2024). https://www.oneusefulthing.org/p/signs-and-portents

Panetta, Kasey. “3 Themes Surface in the 2021 Hype Cycle for Emerging Technologies.” Gartner. (August 23, 2021). https://www.gartner.com/smarterwithgartner/3-themes-surface-in-the-2021-hype-cycle-for-emerging-technologies

Saunders, Cruce. “A New Content Order for the Multi-Channel, Multi-Modal World.” Intercom 64, no. 1 (2018): 9–11. https://www.stc.org/intercom/download/2018-1/

Seymour, Mike, Dan Lovallo, Kai Riemer, Alan R. Dennis, and Lingyao (Ivy) Yuan. “AI with a Human Face.” Harvard Business Review (March-April 2023). https://hbr.org/2023/03/ai-with-a-human-face

Spiers, Elizabeth. “I Finally Figured Out Who ChatGPT Reminds Me Of.” New York Times (January 7, 2024). https://www.nytimes.com/2024/01/07/opinion/chatgpt-generative-ai.html

Wehrle, Frederick T. “UPCEA MEMS 2023 in #Portland was amazing!” [LinkedIn post]. (December 1, 2023). https://www.linkedin.com/posts/frederick-t-wehrle_portland-generativeai-ai-activity-7137128441547558912-KZ8Z


Daniel L. Hocutt serves as Web Manager on the Marketing and Engagement team and teaches as Adjunct Professor of Liberal Arts at the University of Richmond School of Professional and Continuing Studies. His research interests include data analytics and AI in technical communication, digital literacies, and posthuman agency. He’s a research member of the Building Digital Literacy research cluster as part of the Digital Life Institute and serves as Technical Editor for ACM SIGDOC conference proceedings. He’s published in Technical Communication, Communication Design Quarterly, Computers & Composition, Present Tense, and the Journal of User Experience along with several edited collections. dhocutt@richmond.edu

Ann Hill Duin is Professor of Writing Studies and Graduate-Professional Distinguished Teaching Professor at the University of Minnesota where her work focuses on augmentation technologies, digital literacy, and human-AI collaboration. She leads the Building Digital Literacy research cluster as part of the Digital Life Institute (https://www.digitallife.org/), an international research network of multidisciplinary scholars studying the social implications of disruptive digital technologies. Together with Isabel Pedersen, Dr. Duin has published two books focused on AI futures: Writing Futures: Collaborative, Algorithmic, Autonomous (Springer, 2021) and Augmentation Technologies and Artificial Intelligence in Technical Communication: Designing Ethical Futures (Routledge, 2023). ahduin@umn.edu