"The entire process could have been avoided if the client had taken a professional's advice rather than ChatGPT's," colorist Angela Hazelton
tells PEOPLE
Credit: getty
NEED TO KNOW
- New York City-based hair colorist Angela Hazelton says that clients are relying on ChatGPT for their hair color formulations, resulting in mishaps
- Hazelton recalls a customer being “shocked” at her hair color because it didn’t match what the AI tool had predicted
- “The entire process could have been avoided if the client had taken a professional’s advice rather than ChatGPT’s,” the colorist tells PEOPLE
AI tools can be useful, but one colorist is urging salon clients to listen to professionals above all else.
Angela Hazelton of the Marie Robinson Salon in New York City tells PEOPLE that customers have been consulting ChatGPT to find their perfect hair color, but it's backfiring. In one incident, a client asked for a specific gloss formula to remove the warmth from her hair. When the stylist told her that it would also make her golden blonde highlights much darker, the client still insisted on the ChatGPT formula.
"[The colorist] asked the client if she would like her professional opinion and if she would be open to other suggestions and was met with a stern 'no,'" Hazelton recalls. "Her new client wanted the exact formula Chat had come up with to be applied. She even went so far as to ask to see the bottles of color to be sure that was what was being mixed."
The formula was applied per ChatGPT's instructions, followed by a shampoo, hydrating conditioner and gloss. The end result was far from what the client had envisioned.
"To no professional's surprise, the once golden blonde highlights were now a very dark muddy blonde, which might as well be brown when you're used to being sunshine-y blonde," Hazelton says.
The client was "shocked" at her darker color, she recalls, and asked for the gloss to be removed.
"The colorist took her back to the shampoo bowl and clarified her hair, then applied a gentle color remover to remove as much of the pigment as possible," Hazelton says. "The client couldn't believe that Chat would do her dirty like that. What should have been a 30-minute, easy-breezy appointment turned into a two-hour ordeal."
Credit: getty
One of Hazelton's own clients also experienced a ChatGPT mishap. The client had done a color analysis in which she uploaded a photo of herself to the AI tool, and it recommended hair colors that would suit her.
One of the suggestions was a shade of blonde that Hazelton agreed would work well. "The thing that Chat doesn't take into consideration that a professional does is maintenance," she says. "This specific client only wants to color her hair three, maybe four times a year, and in between appointments, she wants the color to grow out seamlessly, no harsh grow-out line."
She continues, "I was extremely transparent about the maintenance this hair color would require. She agreed, and I made her blonde. It complemented her skin tone and eye color beautifully, but 12 weeks after the initial color change, she was sitting back in the chair, asking to return to our original hair color because the blonde upkeep was too much for her."
Never miss a story — sign up for PEOPLE's free daily newsletter to stay up-to-date on the best of what PEOPLE has to offer, from celebrity news to compelling human interest stories.
Hazelton says that ChatGPT can be helpful in the salon, but clients should be aware of its limits.
"Using ChatGPT can be a fun way to explore the possibilities and options you have for hair color or haircut changes," she explains. "It can be used as a tool to make suggestions or even create a mockup photo of what you may look like with a new style. But when it comes to making a concrete decision, I would always defer to a professional."
She continues, "There are so many factors to take into consideration when making a change with your hair, maintenance, integrity of the hair, cost, lifestyle; all things that ChatGPT doesn't think of when it's giving you answers to your beauty questions."
Read the full article here
