ABC Technology Journalist Angharad Yeo discusses the challenges of reporting on Artificial Intelligence.
Australians rely on AI journalism to be informed about the developing technology, but news stories about AI are too often sensationalist, failing to inform the public about critical issues.
In 2022, a Google employee working on an AI called LaMDA was fired after claiming the system had achieved sentience. Media outlets, including The Guardian and The Daily Mail, published the story under alarming headlines, such as the 9 News article ‘Was Google’s AI a sentient being with feelings? This engineer said so, then he got fired‘.
AI experts found the ex-employee’s claims to be unfounded by the evidence. But most of the early media coverage of the story implied that the employee was a whistleblower and failed to weigh the claims against the evidence.
While sensational stories get clicks, they often lack accuracy when it comes to sciences like AI.
Science Communicator Neil De Grasse Tyson highlights the tension between journalism and the rigorous testing required for scientific soundness.
“The urge to get (the story) first means they’re reporting on something that is not yet verified by other scientific experiments, and if it’s not yet verified it’s not there yet.”
“… you’re more likely to write about a story that is more extraordinary, and the more extraordinary it is… the less likely it is to be true. So you need some restraint there, something to buffer the account.” De Grasse says.
Publishing sensationalist stories about AI threatens not only scientific accuracy but also the public interest. Only 12 per cent of Australians feel well informed about how AI is being used, according to a recent study by KPMG and the University of Queensland. As AI continues to transform our lives, this limited understanding highlights the need for reporting on the real threats and benefits of the technology.
Genevieve Bell, Director of 3Ai at ANU, argues the importance of understanding the context of public ideas about AI, often shaped by science-fiction narratives and the ease with which journalism can play into hysteria. Genevieve Bell says journalists need to draw attention to and challenge this context in their reporting.
“We see an over-indexing on fear and anxiety,” she says.
“I would like to see news coverage acknowledge the depths of our imaginations of AI and problematise the ways we talk about both its perils and possibilities today.”
Reporting on sensationalist claims can also distract from the real dangers created by AI technologies. Some AI systems have been shown to have an algorithmic bias against women and minorities, and with the arrival of publicly sourced AI tools like ChatGPT, issues of plagiarism and data security are rampant.
Technology reporter with the ABC Angharad Yeo says that while the danger of AI “taking jobs” is a common media angle, greater focus is needed on the growing issue of data security.
“’Are robots gonna take my job?’ tends to be the space I see media often using as their hook-line, and I don’t necessarily love that because it’s a simple answer. Mechanised systems have been “taking jobs” for many many years, but then we get new ones!”
“But how data can be used against (the public), that is the area that I don’t think there has ever been overblown reporting on because it’s actually way more serious than the average person is wrapping their head around,” Yeo says.
Australians are eager to be better informed about how AI is being used. In the same public trust study, 94 per cent of Australians rated being educated on ‘principles of trustworthy AI’, such as data privacy and security, human agency, and oversight, as the most important to trust in AI systems.
To bridge the gap of understanding, reporting where these principles are being followed and breached should be a priority in AI journalism.
