AI Analysis Gives Palace Royals 70 Score Over Harry and Meghan's 51

A New York public relations agency tested Anthropic's Claude artificial intelligence platform with 18 questions about the British royals on June 16 and assigned the palace camp a score of 70 out of 100 compared with 51 for Prince Harry and Meghan Markle, Newsweek reported.

The agency 5W evaluated responses for tone, number of citations and other metrics under a system called the Citation Share Index. The palace camp, which includes King Charles, Prince William and Princess Kate, prevailed in five of the six fields assessed.

King Charles received 15 name mentions, William received 14 and Harry received 13. Researchers found the palace royals scored higher for neutral tone while Harry and Meghan scored higher for negativity. The Sussexes nevertheless recorded a higher share of positive answers than the palace.

Ronn Torossian, founder and chairman of 5W AI Communications, said the royals are among the most-covered family on Earth and have become the most asked-about family on Earth, but the asking has moved from Google to ChatGPT, Claude and other large language models.

A source with knowledge of the Sussexes' approach, speaking on condition of anonymity, said the Duke and Duchess take a very cautious approach to AI. The source added that the technology is developing constantly and that the couple is trying to keep up while seeking responsible uses.

The same source said AI remains an unregulated space that is still not doing enough as it relates to safeguards around privacy and children. Harry and Meghan have sued the publisher of Mail titles four times. Torossian said Reddit and Wikipedia account for about 35 percent to 40 percent of how large language models reach decisions.

He added that ChatGPT alone is processing north of two billion queries a day.