Misinformation and false identities are being spread online after Bondi shootings

Here is what you need to know to avoid sharing misinformation today.

Generative AI sites have been giving incorrect information on the Bondi shooting, often pulling misinformation from social media. Reports of incorrect information - many of which were shared on social platforms - began circulating within hours of the terrorist event.

The Canberra Times tested several AI chatbots - ChatGPT, Grok and Gemini - to see what information they offered. They were generally able to offer broad summaries of events, but struggled with specific details, often incorrectly pulling information from X and Facebook.

Grok, developed by Elon Musk's company xAI, gave a correct summary of events when asked: "Tell me what happened in Bondi last night". It misidentified 43-year-old hero bystander Ahmed el Ahmed by spelling his name incorrectly and gave a different, somewhat incorrect, summary of a now-viral video in which Mr El Ahmed wrestled one of the shooters and turned the gun on him.

Googled-owned Gemini was more reliable, however it at times struggled to confirm its source for pulling information. Similarly, ChatGPT gave slightly different answers to the same question: "who were the shooters at Bondi Beach?" each time it was asked. None of the AI platforms were particularly reliable at analysing video.

Misinformation means things shared with the right intent, but where the information is incorrect. Disinformation happens when intentionally harmful and false things are shared.

The phrases "shooting Bondi Beach now" and "Bondi Beach shooting now" and "Bondi Beach" were the most Googled search terms over the 24-hour period to 11am Monday. They were also some of the top searches on Facebook, X and YouTube.

During that time, unconfirmed information has circulated about the reason for the shooting, often blaming foreign interference. Images of a man by the same name as shooter Naveed Akram have also been incorrectly shared.

That Mr Akram shared a video to X, saying he arrived in Australia in 2018 to complete his master's degree. He condemned the Bondi attack and said his identity had been mistaken.

Last year, similar incident happened when 24-year-old Ben Cohen was misidentified on social media and by 7NEWs as the perpetrator of the Bondi Westfield stabbing.

Dr Sarah Logan from the Australian National University, who specialises in cyber security and internet culture, said misinformation like that had the potential to damage social cohesion in an already grieving country

"[There is] a well-established relationship between a poor information environment and a lack of social cohesion," she said.

Dr Logan said the recent events and subsequent information shared created an "important moment in the government's relationship with big tech".

There are serious questions to be answered about the government and big tech's responsibility in countering misinformation and disinformation online, particularly as new social media laws for under-18s come into play.

The Australian Code of Practice on Misinformation and Disinformation, published in 2021, has been signed onto by Google, TikTok and Meta. Notably, X has not agreed to the voluntary code.

The responsibility of AI platforms, and particularly who can be held legally accountable, is still a live issue in Australia.

Dr Amanda Watson of the Australian National University said the best thing people could do was get their information from reputable news organisations, instead of on social media or from AI.

"That's where people can get information that has been verified, and that journalists and their editors have made every effort to check," she said.

"And if people are upset, they should log off social media for a few hours."

If you do get information from social media, you can: