AI in Business
How AI-generated plain English changes investor interest in mutual funds
A new study reveals that AI-generated, plain-English summaries of mutual fund prospectuses significantly increase retail investors' comprehension and willingness to invest. The preference for simple language holds true for both financial novices and highly experienced investors.
How workers are actually using generative AI on the job
A new survey of the US labor market reveals that generative AI adoption is widespread across many professions but shallow within them, heavily dependent on a worker's personal willingness to learn the technology.
How generative AI is changing the volume and tone of consumer reviews
When Italy briefly banned ChatGPT in 2023, researchers used the four-week outage to measure how AI tools change online reviews. The result: fewer, shorter, less opinionated hotel reviews when the tool went away.
Can AI Design a Better Ad? A New Study Puts It to the Test
Researchers fine-tuned an AI image generator on real consumer reactions and found its ads often matched or beat human-made ones, except when a product needed to stand out through surprise.
Why salaried jobs might handle AI differently than hourly ones
A new working paper models what happens when workers enjoy their tasks enough to work on them unpaid. The result reshapes how firms choose what to automate and warns that payroll data alone can obscure very different jobs.
Adding a human to oversee AI doesn’t automatically build trust, study finds
A study of AI-generated investment reports finds that labeling content as "human-supervised" can lower trust. What builds credibility is a subtler inference: the sense that someone competent actually reviewed the work before you saw it.
Generative AI in B2B sales: A study of five companies finds a clear division of labor
A new study distinguishes between two kinds of sales creativity: the recombinational work generative AI can enhance, and the contextual judgment that stays fundamentally human.
The generative AI apology: better than a human’s, sometimes
New research finds generative AI outperforms human agents at handling minor service failures, but loses its edge when initial fixes fail, unless empathy or compensation is added.
New research shows exactly where virtual avatars fall short of human spokespeople
Virtual influencers can sell excitement about as well as humans can. But when the emotion is gratitude, hope, or pride, digital endorsers fall short, unless audiences already share their values.
The pronoun trick that makes virtual influencers feel more human
New research finds that virtual influencers seem more human, and more persuasive, when they address audiences directly with "you" instead of narrating personal stories with "I."