At the 20th annual MozCon, Britney Muller, founder of Data Sci 101, gave an eye-opening presentation on AI and its impact on digital marketing.
Her session, “The Hidden Side of AI: What Marketers Need to Know,” provided a comprehensive overview of AI’s current and future potential.
Muller discussed the ethical considerations, practical applications and limitations of AI, and offered valuable advice for marketers.
The Emergence of Generative AI
Mueller began by discussing the rise of generative AI, which sits at the intersection of AI, machine learning, deep learning, and natural language processing (NLP).
She explained:
“Generative AI in particular has emerged from this interesting overlap of fields.
“We have AI that incorporates machine learning. Within machine learning, we have deep learning. And then we have NLP, or natural language processing, which leverages human language.”
Photo taken by the author at MozCon in June 2024.A large part of Mueller’s presentation focused on the critical role of training data in AI models.
She emphasized:
“I've said before that AI reflects its training data, but I'll take that idea even further: AI augments its training data.”
Muller highlighted that datasets like Wikipedia lack diversity because most contributors are male, which can perpetuate bias in AI output.
Photo taken by the author at MozCon in June 2024.Practical Applications and Limitations of AI in Marketing
Gen AI's Areas of Expertise
Mueller presented a wide range of practical applications of AI in marketing, as shown in one of his slides.
Photo taken by the author at MozCon in June 2024.She explained:
“LLMs are generally good at all of these things, but it's an unpopular opinion that content generation is one of their weakest competencies. They are much better at sentiment analysis, labeling as categories, and providing code support.”
Additionally, she shared slides showcasing specific applications of GenAI SEO/Marketing, including:
- Auto title and meta description
- Atta Cleaning
- Code Assistance
- Accelerate creativity and ideation
- Personalized Outreach
- Sentiment Analysis
- Content Improvements
- Chatbots
- Transcription of meeting notes
Photo taken by the author at MozCon in June 2024.What GenAI is bad at
Photo taken by the author at MozCon in June 2024.Muller explained the limitations of LLMs, who struggle with tasks such as:
- Factual accuracy
- Common sense reasoning
- Understanding the context
- Dealing with unusual scenarios
- Emotional Intelligence
- Mathematics/Calculations
Marketers need to be aware of these pros and cons when incorporating AI into their strategies.
Rapid Engineering Tips
To help marketers take advantage of generative AI, Mueller offered practical tips for rapid engineering.
Photo taken by the author at MozCon in June 2024.Her three suggestions were:
- Explain a task the same way you would explain it to a person
- Explain what you want with an example
- Give your models a “role” and communicate your target audience
She gave this advice:
“You explain the task or problem just like you would explain it to a person. There's been a lot of research on prompt engineering, and they say, oh, this works, this doesn't work. The biggest lesson from all of this research is examples. Just show them a model and say this is good or bad, and this is what the output should look like.”
Muller shared slides of generative AI tools and resources, including Colab, Kaggle, GPT for Sheets, Ollama, WordCrafter.ai, and her own DataSci101.com.
Photo taken by the author at MozCon in June 2024.Key Points and Future of AI in Marketing
Mueller concluded his presentation with some key points summarized in his final slide.
Photo taken by the author at MozCon in June 2024.She stressed the need to take a human-centric approach, recognising AI’s potential as an assistive technology rather than a complete replacement for human expertise.
The main points are:
- GenAI is a predictive technology
- A model is only as good as its training data
- Marketers have the power to dream up the next great GenAI application
- Be present in online places where conversations about your products and services are happening
She said:
“We need to have more conversations about human-centric AI. What are the best models to support the people we work with? And this is a predictive technology. The models are only as good as the training data, so it's a complementary technology. It's not and never will be a complete replacement for humans.”
In summary
Mueller’s insights serve as a valuable guide for navigating the complex world of AI.
Throughout his presentation, Müller reiterated that AI should be seen as an assistive technology, not a complete replacement for human expertise.
She encouraged marketers to identify tasks that AI can help accelerate or automate while still retaining the human touch.
Mueller's main messages to marketers are to adhere to ethical practices, prioritize human needs, and leverage AI's strengths while recognizing its weaknesses.
Featured image: abnalladin/Shutterstock
