As generative tools spread through offices and classrooms, veteran executive Sol Rashidi is urging users to keep their minds in the loop. Rashidi, who has spent 15 years working in artificial intelligence, cautions that convenience should not replace judgment. Her message arrives as organizations race to deploy AI for writing, coding, research, and decision support, and as educators rethink how students learn in a world filled with instant answers.
The concern is simple: when software drafts, summarizes, and recommends, people may stop reasoning for themselves. That risk has broad implications for skills, trust, and accountability. It also shapes how companies train workers and how schools teach critical thinking.
A Veteran’s Warning
“It’s important to make a conscious effort not to outsource your thinking.” — Sol Rashidi
Rashidi’s point is not to reject AI. It is to keep agency. Her experience across corporate AI programs gives weight to a growing debate about when to rely on machine help and when to slow down. The line is not always clear. Many tools now produce polished text, code, and analysis. That output can be useful, but it can also be wrong, biased, or shallow.
Her view aligns with a broader industry shift. Many leaders now stress “human in the loop” oversight and source checking. The idea is to use AI for speed while preserving human judgment.
Background: Rapid Adoption Meets Old Lessons
AI assistants moved from experiments to daily use in a short time. Workers draft emails with chatbots. Developers use code completion. Students review study guides generated on demand. The gains are real: faster drafts, quicker prototypes, and easier summaries.
History offers a caution. Automation often brings “automation bias,” where people trust a system even when it is wrong. Aviation, health care, and finance introduced checks and training to counter that bias. The same playbook applies to AI writing and analysis tools. Verification and accountability must keep pace with adoption.
Researchers and policy groups have also flagged risks from overreliance. These include hallucinated facts, uneven data quality, and hidden training gaps. While many reports highlight productivity gains, they also warn about skill atrophy if people stop practicing core tasks.
Rising Use in Work and School
Companies are setting rules for when and how to use AI. Many require disclosure in client-facing work and documentation of sources. Some restrict the use of sensitive data in third-party systems. Others are building internal tools with guardrails and audit logs.
Classrooms face a separate challenge. Educators want students to learn how to use AI without letting it do the learning for them. That means assignments that reward reasoning, show steps, and require citations. It also means teaching how to question a confident answer and how to compare outputs with trusted sources.
- Skill risks: Overreliance can weaken writing, coding, and research habits.
- Quality risks: Fluent answers may include errors or bias.
- Privacy risks: Sensitive data can leak if entered into public tools.
Balancing Speed and Judgment
Experts suggest simple practices to keep thinking active. Start with a clear question or hypothesis. Use AI to explore options, not to dictate conclusions. Cross-check claims with primary sources. Track what the system added and what the human decided.
Teams can formalize this approach. Set review steps for high-stakes work. Define when to escalate to a human expert. Log prompts and outputs for learning and oversight. Encourage employees to explain their reasoning, not just paste results.
These habits help separate drafting from deciding. They also give managers a view of how AI shapes work quality and speed.
Competing Views on AI’s Role
Some practitioners argue that AI expands thinking by removing busywork. They say that when software handles summaries and boilerplate, people can focus on strategy and creativity. Studies on productivity often support this view for routine tasks.
Others warn that shortcuts can weaken core skills. They point to students who skip the struggle that builds understanding and to teams that accept the first fluent answer. Both groups agree on one point: transparency and verification matter.
What Comes Next
Enterprises are likely to invest in training that pairs tool use with reasoning skills. Schools will update curricula to teach prompt design, source review, and citation. Regulators may push for clearer disclosures on AI-assisted content. Vendors will improve controls, but users will still need judgment.
Rashidi’s advice offers a simple guide for this moment. AI can draft, suggest, and summarize. People must still decide. The next phase of adoption will reward those who move fast without surrendering their thinking.
The takeaway is clear: use AI for leverage, not for autopilot. Expect more organizations to set rules that keep humans accountable, and watch for training programs that make critical thinking a core skill again.
