AI at Work
Harvard Business School faculty research shows how artificial intelligence can drive experimental agility and innovation, but also reveals reputational risks, data quality pitfalls, and the potential for bias.
What does HBS faculty research say about AI in the workplace?
🤝 Human-AI teamwork leads to more innovation. When humans leverage AI, they can identify innovative ideas for addressing social problems more efficiently than either humans or AI could on their own.
🎯 Success with AI depends on the user’s skills and input. Entering faulty data leads to suboptimal plans.
🤔 AI can seem right even when it’s wrong. Tools can be persuasive enough to convince users to question their own conclusions.
📈 Productivity gains can help managers focus on what matters. A study of software coders found that AI didn’t just help them code more efficiently, it empowered them to approach their jobs differently and follow their interests.
👩💼 AI could widen the gender pay gap. Women are adopting AI tools at a 25 percent lower rate than men, potentially putting them at a career disadvantage.
🗣️ Communication generated with AI can face credibility questions. “Algorithm aversion” is real, and constant AI-generated outreach might undermine a leader’s message.
🛡️ AI can help combat cyberattacks. Tools can help leaders with crisis scenario planning, simulations, and decision-making.
💬 Chatbots can help people sound more human. AI helped human agents respond to chats some 20 percent faster, and the technology helped people reply with more empathy and thoroughness.
🧩 Algorithms can’t do everything. Large language models are as reliable as the data they’ve been trained on, and success requires iteration. Research suggests they may have a long way to go before they can truly replicate this near-instant flexibility that is typically second nature for humans
🧠 Soft skills still matter. The ability to communicate, interact, and think critically will underpin how people acquire more advanced professional skills, even in an AI world.
What do companies need to succeed with AI?
Well-trained talent who can deploy AI thoughtfully.
AI tools that can scale with an organization’s ambitions.
Processes that ensure strong data quality.
Adaptable cultures that value experimentation and problem-solving.
Leaders with strong understanding of AI and good judgment.
Knowledge of AI's ethical concerns and potential for bias and “hallucinations.”
What are some potential missteps with AI?
Relying on the tools excessively, which can erode the creative process.
Scaling too quickly—AI continues to evolve and improve, but still has limits.
Deploying AI without the right cybersecurity protections.
Accepting AI output without testing or questioning its results.
