Navigating the Pros and Cons of GPT-Powered AI

Brittany Clausen
Aug 30
4 min read

The launch of OpenAI’s GPT-5 has reignited conversations about the potential and limitations of large language models. While the new version brings noticeable improvements—sharper reasoning, smoother coding, and fewer “hallucinations”—reviews have been mixed. Some celebrate it as a leap forward, while others view it as an incremental step rather than a groundbreaking shift (SAN).

This duality highlights a truth we can’t ignore: AI, even in its most advanced forms, is a tool. Its effectiveness depends on how we use it, what safeguards we implement, and how mindful we are of its broader impacts.

The Benefits of GPT AI

Improved Reliability
GPT-5 has reduced errors in logic and factual accuracy, making it a stronger companion for professionals. Developers have found it more cost-efficient in coding tasks compared to competitors like Claude or Gemini (Wired).
Stronger Professional Applications
Business users note better performance in persuasive and corporate writing. According to Tom’s Guide, GPT-5 is particularly adept at clarity and structure in workplace communication.
Accessibility and Speed
For organizations under tight deadlines, GPT tools can quickly draft policies, generate training outlines, or simulate communication styles—reducing workload while expanding creative options.

The Challenges and Limitations

Incremental Progress
While GPT-5 is better, it’s not transformative. As Financial Times points out, AI innovation may be slowing, with returns on new versions looking more modest than early breakthroughs.
Variable Creativity
Critics note that GPT-5’s prose is polished but sometimes uninspired. For teams relying on creativity, it can produce “safe” but formulaic results.
Hallucinations Persist
AI can still confidently generate incorrect or misleading content. Over-reliance without verification could lead to costly errors.
Intellectual Property & Security Risks
Sharing sensitive or proprietary data with AI models can create security exposures. Outputs may also raise copyright and IP questions, especially in industries producing creative or original work.

The Environmental and Social Costs

Beyond productivity, AI has a hidden cost: energy consumption. Every query requires computing power, and as use scales, the environmental footprint grows. A Washington Post analysis shows AI data centers are major energy guzzlers, straining local power grids and water supplies.

This burden often falls disproportionately on marginalized communities where data centers are built. For example, in Memphis, residents are fighting the environmental impact of new AI infrastructure backed by Elon Musk’s xAI (TIME). Teen Vogue frames this as a continuation of environmental racism, where Black and Brown communities disproportionately absorb the costs of technological expansion while reaping fewer benefits.

As AI expands, organizations must not only weigh productivity gains but also advocate for ethical and equitable practices in its development.

What You Need to Watch Out For

For leaders, the responsibility goes further: you must safeguard how your people use AI. Here's how you can do that:

Data Security & Confidentiality
Employees may unintentionally paste confidential data into GPT, risking leaks and compliance violations. Draft a clear AI-use policy banning sensitive data uploads.
Equity & Environmental Impact
Every query consumes energy, often straining marginalized communities. Ask AI vendors tough questions about sustainability and equity.
Employee Skill Atrophy
Heavy reliance on GPT can weaken critical thinking and communication skills. Position GPT as a draft assistant, never the final decision-maker.

Legal & IP Risk
Outputs may contain copyrighted or plagiarized material. Train your teams to vet outputs before publishing or external use.
Bias in Outputs
GPT reflects biases in its training data, which can surface in hiring, communications, or training content. Encourage teams to critically review outputs for fairness and inclusivity.

Turn the Lightbulb On: Use AI With Intention

Generative AI is powerful, but it isn’t neutral. “Turning the lightbulb on” means moving from autopilot to accountable, intentional use—pausing before you copy/paste, ask one smarter question, and take ownership of the final output.

For everyday users, do a quick LIGHT check (30 seconds):

L — Learn the source: Double-check facts, dates, and claims. If you can’t cite it, don’t ship it.
I — Inspect for bias: Scan for stereotypes, exclusionary language, or one-sided framing.
G — Guard confidentiality: Never paste sensitive data (PII, contracts, salaries, PHI).
H — Humanize the voice: Edit for your tone, context, and audience. Avoid that generic “AI gloss.”
T — Timebox usage: Cap your AI drafting time; then switch to human thinking. Overuse = dependency.

Why this matters:

GPT still hallucinates—unchecked content can spread misinformation.
Overuse can flatten your voice into a polished but generic tone.
Hours in AI can create dependency, dulling creativity and judgment.

Micro-habit: Before you hit send, ask: What did AI get me started on—and what value did I add as the human in the loop?

Final Thought

GPT-5 demonstrates the promise and pitfalls of generative AI. Its advancements show how far we’ve come in using machines to support human work—but its limitations remind us that AI is not a substitute for human wisdom, empathy, or accountability.

The best way forward? Use GPT as a tool to extend human potential, not as a crutch that replaces it. Leaders must set boundaries, safeguard data, and protect culture to ensure AI strengthens—not weakens—our people and organizations. Keep the lightbulb on. 💡

References

Financial Times. (2025). Is AI hitting a wall? Financial Times. https://www.ft.com/content/d01290c9-cc92-4c1f-bd70-ac332cd40f94

SAN. (2025). OpenAI’s GPT-5 launch draws mixed reviews despite major advances. San.com. https://san.com/cc/openais-gpt-5-launch-draws-mixed-reviews-despite-major-advances/

Teen Vogue. (2025). ChatGPT is everywhere. The environmental costs are staggering. Teen Vogue. https://www.teenvogue.com/story/chatgpt-is-everywhere-environmental-costs-oped

TIME. (2025). Inside the Memphis community battling Elon Musk’s xAI data center. Time. https://time.com/7308925/elon-musk-memphis-ai-data-center/

Tom’s Guide. (2025). I tested ChatGPT-5’s writing abilities across 3 challenges. Tom’s Guide. https://www.tomsguide.com/ai/chatgpt/i-tested-chatgpt-5s-writing-abilities-across-3-challenges-heres-how-it-performed

Washington Post. (2025). ChatGPT is an energy guzzler. These things you’re doing are worse. The Washington Post. https://www.washingtonpost.com/climate-environment/2025/08/26/ai-climate-costs-efficiency/

Wired. (2025). GPT-5 for software engineering review. Wired. https://www.wired.com/story/gpt-5-coding-review-software-engineering/

Tech Learning. (2025). GPT-5 has launched: Here’s everything teachers need to know. Tech Learning. https://www.techlearning.com/news/gpt-5-has-launched-heres-everything-teachers-need-to-know