Brand of Brothers: the branding & marketing podcast podcast

AI Ethics: Data Usage & Privacy

0:00
32:37
Rewind 15 seconds
Fast Forward 15 seconds

Welcome back to Brand of Brothers with Doug Berger and Johnny Diggz, where branding, marketing, business, creativity, and emerging technology get unpacked with clarity, humor, and zero fluff. In this episode, we continue our conversation about the Higgins-Berger Scale of Generative AI Ethics, focusing on data usage and privacy: where AI training data comes from, what happens to uploaded work, and how businesses can use AI without exposing confidential information or violating creators’ rights.

Before an AI system can generate a word, image, song, video, design, or idea, it has to learn from data. That raises a foundational question:

When creators share their work online, are they only giving people permission to see it, or are they also giving AI companies permission to train on it?

 In this episode:
• What “data usage and privacy” means within the Higgins-Berger Scale
• The difference between owning your work and licensing a platform to use it
• Why public availability is not necessarily informed consent for AI training
• How social media terms may give platforms broad rights to process uploaded content
• Why copyright, licensing, privacy, consent, and provenance overlap without meaning the same thing
• The ethical difference between licensed, compensated, openly licensed, public-domain, and nonconsensually collected material
• Why Adobe Firefly may score more favorably because of its licensed and public-domain training approach
• Why responsibly sourced training data does not eliminate every risk in the generated result
• How AI can still produce celebrity likenesses, recognizable trademarks, protected designs, or patented visual effects
• The cake analogy for understanding hosted AI, locally operated models, and models built from known data
• Why hosted systems resemble eating restaurant cake without knowing every ingredient
• How running a downloaded model locally can improve privacy without resolving uncertainty about its training data
• Why building from documented, properly licensed data offers greater provenance and control
• Why privacy is about understanding and controlling how information will be used, not simply keeping it secret
• The risks of uploading client strategy, brand launches, internal evaluations, goals, or competitive information
• Why commercial AI privacy promises still require governance, judgment, and accountability
• How third-party photography, fonts, illustrations, templates, and other licensed assets complicate private AI training
• How the 80/20 principle applies to AI-assisted creative workflows
• Why the final 20% still requires human judgment, taste, expertise, correction, and craft
• Why AI is often more defensible for ideation, analysis, and internal processes than as the final creative source
• The difference between asking “Can we legally do this?” and “Should we ethically do this?”
• Why the HBS is a scale for navigating gray areas, not a simple ethical-or-unethical verdict
• Whether requiring creators to opt in to AI training would protect rights or restrict innovation
• A preview of the next HBS conversation about displacement impact and human labor

Whether you are an agency owner, creative director, designer, writer, musician, content creator, entrepreneur, CMO, technology leader, or business owner exploring generative AI, this episode offers a practical way to evaluate training data, privacy, consent, provenance, intellectual property, and professional responsibility.

Listen now to learn how to:
• Evaluate data usage and privacy in an AI-assisted workflow
• Separate public access from meaningful consent
• Distinguish local processing from ethical training-data provenance
• Identify confidential information that should not be casually uploaded to an AI platform
• Decide when AI should remain assistive rather than become the final creative source

Explore the Higgins-Berger Scale:
• Interactive HBS Utility: https://r3mx.com/hbs-interactive-utility/
• Full Higgins-Berger Scale Article: https://r3mx.com/the-higgins-berger-scale/

Ethically Sourced and Rights-Respecting AI Resources:
• Fairly Trained Certified Models: https://www.fairlytrained.org/certified-models
• KL3M, the first Fairly Trained-certified LLM: https://273ventures.com/kl3m-is-first-fairly-trained-llm/
• Common Pile and Comma Models: https://huggingface.co/common-pile
• Adobe Firefly’s Training Approach: https://www.adobe.com/ai/overview/firefly/gen-ai-approach.html

Presented by REMIXED, the full-service branding agency helping companies craft, launch, and grow powerful brands.

 Music by PRO
 Visit us at BrandShowLive.com
 Follow along at @BrandShowLive on all socials
 Like, subscribe, and share to support the show and keep the conversations going

More episodes from "Brand of Brothers: the branding & marketing podcast"