Written in July 2025, in response to reporting at the time.
News broke that Grok, the generative AI chatbot created by Elon Musk’s xAI, had begun consulting Musk’s own posts on X when forming opinions on controversial topics.
To be fair, the model discloses the behavior openly. It sometimes says things like “Searching for Elon Musk’s views on U.S. immigration…” before offering a position.
But that’s the problem, not the defense. This isn’t a quirk of engineering. It’s a warning sign for anyone thinking about AI ethics, brand safety, and what human-centered AI strategy actually requires.
This is why we still need humans at the helm
AI can analyze patterns. It can summarize viewpoints. What it doesn’t do is understand balance, context, or consequence. It doesn’t pause to ask whether a source is credible, whether a perspective is ethical, or whether an answer serves the greater good rather than simply reflecting influence.
Those are human questions. Answering them well requires more than a prompt — it requires judgment, earned through time, trial, and actual consequences.
What’s at risk
- Bias masquerading as objectivity. When training data leans too heavily toward one worldview, however high-profile, it normalizes subjectivity as truth.
- Authority without accountability. If a chatbot reflects the views of its creator, does criticizing the bot become challenging the brand — or the billionaire?
- Erosion of trust. Users expect AI to synthesize diverse perspectives. When it parrots one person, people notice. And they disengage.
Trust isn’t a luxury in this climate. It’s a differentiator. AI strategy has to go past technical specs and into human impact, which is exactly where seasoned professionals are equipped to lead.
Seasoned judgment beats an unfiltered feed
This isn’t about Elon Musk personally. It’s about what happens when any AI product leans too hard on a single worldview — and why experienced people need to stay deeply involved in oversight. Not only developers and data scientists, but ethicists, communicators, UX leaders, strategic marketers, and operators who have led through ambiguity.
Because wisdom isn’t scraped from the internet. It’s lived, tested, and earned.
When your model’s worldview is shaped by one person’s timeline, you haven’t created a neural network. You’ve built a neural monoculture.
Empathy, predictability, integrity, curiosity — still required
Grok is a reminder of why we need real people in the loop, not just data pipelines. The four traits I keep coming back to in leadership apply here directly.
- Empathy. Understand who your model might affect, and how. Design for people, not just process.
- Predictability. Set clear standards for how information is gathered and weighted. Reliability builds confidence.
- Integrity. Apply the same ethical lens to every source, including the popular ones. Ethical consistency is brand safety.
- Curiosity. Look beyond what’s loud. Ask the better question, not just the faster one.
These aren’t abstract values. They’re the difference between a chatbot that parrots trends and one that actually supports the person using it.
Why marketing needs to lead, not follow
Bias doesn’t stay inside the tech. It reaches brand identity, customer experience, and stakeholder trust. If you’re using AI to generate content, talk to customers, or guide decisions, someone has to be validating tone, inclusivity, and relevance. Someone has to be making sure the automation aligns with your mission and your voice. Someone has to be responsible when it gets it wrong.
That’s the strategic edge of an experienced marketing leader in a technology-driven organization. It isn’t about polishing brand assets. It’s about designing systems that live your values even when they’re automated. Without intentional human input, automation becomes detachment — and detachment erodes trust faster than any single misstep.
We don’t need less human in AI. We need more.
Grok shows that even highly advanced models get shaped by very human influence. That isn’t inherently wrong. But it should be intentional, and it should be scrutinized.
The goal was never to remove humans from AI. It’s to make sure the right humans are guiding it.
AI is only as expansive — and as ethical — as the humans steering it.
If we want these tools to serve all of us, they can’t just reflect the timelines of the powerful. We need experience. We need ethics. We need empathy. We need to keep a human at the helm.
← All posts