The use of artificial intelligence in journalism to create content is seen by media professionals as both an opportunity and a threat, according to a study released today.

The study, released by the London School of Economics’ JournalismAI program and supported by Google News, surveyed members of more than 100 media outlets in 46 countries from April to July.

Almost three-quarters (73%) of respondents said that genetic artificial intelligence, i.e. tools such as ChatGPT or Google Bard that are capable of generating texts, represent “new opportunities for journalism, improving efficiency, productivity and creativity”. 85% of respondents have tried AI for writing summaries or headlines.

But more than 60% also expressed concerns about the ethical implications of artificial intelligence, as well as the quality of texts, accuracy, balance and transparency.

“Journalism around the world is going through a new period of technological change that both excites and frightens,” Charlie Beckett, who co-authored the research and is the program’s director, said in a statement.

The study shows that genetic artificial intelligence is both “a potential threat to the integrity of information and media,” but that these tools also offer “an incredible opportunity to make journalism more effective and credible,” he continued. .

Participants in the study highlight, for example, the benefits of artificial intelligence for transcribing interviews and highlight the need for the content created in this way to be verified to avoid “biases and inaccuracies”.

This study showed, among other things, the difficulties faced by the countries of the South. “Artificial intelligence technologies are mainly available in English and not in many Asian languages. We need to rush in two directions, to create artificial intelligence systems and to make them work in our languages”, underlined one of the respondents who comes from the Philippines.

According to Mira Yasin, who co-authored the study, the benefits of artificial intelligence are concentrated in the countries of the North, the harmful effects in those of the South.

“If we are serious about benefiting from AI in an equitable way, it is imperative” that we take a power-conscious approach to its development and adoption.