Efficiency, ethics and impact: staying ahead of AI

by Yazan Radaideh:

Introduction
The communications industry is hurtling toward an AI-dominated future, and frankly, we’re not ready. Tools like ChatGPT and sentiment analyzers promise efficiency, but their reckless adoption risks eroding public trust and amplifying systemic biases. Let’s be clear: AI isn’t just a tool—it’s a revolution that demands ethical guardrails. As communicators, we’re not bystanders here. We’re the ones who’ll decide whether this technology becomes a force for inclusivity or a weapon of mass manipulation.

This isn’t a neutral analysis. It’s a rallying cry. Below, I break down AI’s transformative role in press releases, sentiment analysis, and chatbots, but with a critical lens on the ethical landmines we’re ignoring. Spoiler: If you’re not questioning AI’s biases or demanding transparency, you’re part of the problem.

1. Automated Press Releases: Efficiency at the Cost of Authenticity?

The Good (Because We Need to Acknowledge It)
Yes, AI can churn out a press release in seconds. Tools like Jasper and PR Newswire are undeniably useful for repurposing earnings reports or localizing content. But let’s stop pretending they’re flawless. I’ve seen these tools butcher brand voice, misrepresent nuanced announcements, and prioritize SEO over substance.

The Ugly Truth About “Speed”
Take the Associated Press’s 2014 experiment with Automated Insights. While scaling from 300 to 4,400 earnings reports sounds impressive, what’s unsaid is how many human editors were left cleaning up AI’s dry, formulaic output. Speed means nothing if the result lacks humanity—or worse, contains errors that slip through cracks in oversight.

My Unpopular Opinion
AI-generated press releases should come with disclaimers. Period. If we’re using these tools, we owe audiences transparency. And if your CEO balks at adding “Created with AI assistance” to a footer, ask them: What are you hiding?

2. Sentiment Analysis: The Bias Machine We’re All Ignoring

Why It’s Dangerous to Outsource Empathy
Sentiment analysis tools claim to decode public perception, but they’re often tone-deaf. Take Brandwatch or Talkwalker: these platforms routinely misclassify sarcasm, dismiss cultural nuance, and—most damningly—pathologize marginalized dialects. A 2022 study found that African American Vernacular English (AAVE) is flagged as “negative” 30% more often than standard English. That’s not a glitch—it’s a reflection of who’s not in the room when AI models are built.

Starbucks’ Pumpkin Spice Latte Problem
Starbucks touts its AI-driven sentiment analysis for boosting revenue, but let’s dissect this. By hyper-optimizing campaigns for positive sentiment, are we silencing valid criticism? When regional feedback shapes menu offerings, who gets left out? Rural communities? Non-English speakers? Sentiment analysis isn’t neutral—it’s a filter that privileges majority voices.

What No One Wants to Admit
If your sentiment analysis tool hasn’t been audited for racial, gender, or cultural bias, you’re not just lazy—you’re complicit. And no, adding a “human moderator” as an afterthought isn’t enough.

3. Chatbots: The Illusion of Connection

The KLM Fairy Tale
KLM’s chatbot handles 60% of customer inquiries. Sounds great, right? Until you realize that “efficiency” often means deflecting complex issues to FAQs or shutting down frustrated users with scripted replies. Chatbots like Sephora’s Virtual Artist might boost conversions, but they’re also normalizing a world where genuine human interaction is a premium service.

When Bots Become Bullies
The National Eating Disorder Association (NEDA) learned this the hard way. Their AI chatbot, designed to support vulnerable users, ended up doling out toxic diet advice. This wasn’t an anomaly—it’s what happens when we prioritize cost-cutting over compassion.

My Take: Burn the Scripts
Chatbots should be a last resort, not a first response. If your user can’t reach a human within three clicks, you’ve failed. And if you’re not petrified of chatbots collecting sensitive data, you haven’t read enough GDPR horror stories.

4. Ethical AI Isn’t a Buzzword—It’s Survival

The Four Commandments (Because “Guidelines” Are Too Weak)

1. Transparency or Bust: Label every AI-generated output. If you’re ashamed to admit AI wrote it, maybe it shouldn’t exist.
2. Bias Audits, Not Checkboxes: Hire third parties to tear your models apart. If your team lacks diversity, your audits are theater.
3. Human Oversight ≠ Human Theater: “Reviewing” AI outputs isn’t enough. Rewrite them. Every. Single. Time.
4. Regulate Before It’s Too Late: The EU’s AI Act is a start, but complacent companies will wait for lawsuits to act. Don’t be one.

Conclusion: Choose Your Side

AI won’t “disrupt” communications—it already has. The question is: Will you wield it as a tool for equity, or let it amplify existing biases? Will you prioritize speed over truth, algorithms over accountability?

Here’s my challenge to you: Fight for the human edge. Use AI to draft, but never to decide. Deploy sentiment analysis, but audit it ruthlessly. Build chatbots, but never without an escape hatch to a living, breathing person.

The future of communication isn’t just about staying ahead—it’s about not losing our souls in the process.

+++

Yazan Radaideh is a seasoned PR & Communications Specialist with over 12 years of experience crafting compelling narratives and driving impactful communication strategies. His expertise spans public relations, media outreach, and strategic communications, with a proven track record of elevating brand visibility and shaping public perception. He is a Strategic Columnist and a #WeLeadComms honoree, and is based in Amman, Jordan.

Written by: Editor