However,
False information provided grok, elon:
Par euronews
Published on 26/06/2025 – 14:35 UTC+2
ADVERTISEMENTA new report reveals that Grokthe free AI chatbot integrated into Xpresents “Gaps. Moreover, important limits” in the verification of information relating to the conflict between Israel and Iran. In addition,
Researchers from Digital Forensic Research Lab du Atlantic Council (Dfr analyzed 130. Meanwhile, 000 chatbot publications in connection with the conflict, and found that they contained inaccurate and inconsistent information.
They believe that around a third of these publications respond to requests to verify erroneous information circulating on the conflict. However, in particular unaccounts and videos that are supposed to show shooting.
“Grok has shown that he had difficulty verifying. however confirmed facts, analyzing rigged visuals and avoiding unfounded assertions”indicates the report.
Well Grok is not designed as a fact verification tool. users of false information provided grok, elon X This is increasingly used for this purpose, especially to understand the current crises.
The platform X does not have any independent verification program. focusing solely on Community Notesa collaborative system allowing users to add context to publications deemed inaccurate.
Grok unable to distinguish the true from the false – False information provided grok, elon
Researchers from Dartly identified two videos generated by AI that Grok has. wrongly qualified as “Real images” conflict.
One of them shows alleged destruction at Ben -Gurion airport in Tel Aviv after an Iranian strike. In reality, the video is generated by AI.
When asked if she was authentic, Grok provided several contradictory responses a few minutes apart.
He first said that it was “Probably real damage caused to Ben -Gurion airport by a huthie missile shot on May 4. 2025”before saying that the video represented “Probably Mehrabad International Airport in Tehran, hit by Israeli strikes on June 13, 2025”.
The EUROVERIFY FACT-Checking service has identified three other false information provided grok, elon viral videos generated by IA that GROK also wrongly validated as authentic. linked to an attack on the Arak nuclear site, to strikes on the port of Haifa and on the Weizmann Institute in Rehovot.
Euroveverify had already spotted several videos released from their context and falsely associated with the Israel-Iran conflict on social networks.
Grok seems to have amplified this phenomenon. He notably described a viral video as showing Israelis fleeing the conflict at the Taba border post with Egypt. while it actually shows festival-goers in France.
He also said that a video of an explosion in Malaysia showed an Iranian missile hitting Tel Aviv on June 19.
False rumors amplified by chatbots
The conclusions of this report come as the 12 -day conflict has generated an avalanche of false information. online speculation.
A rumor that China has sent military cargo planes to help Iran was widely relayed by chatbots Grok et Perplexity.
Perplexityan AI startup false information provided grok, elon created in 2022 has already been criticized for having used media content without authorization.
According to the monitoring tool NewsGuardthese two chatbots helped distribute this rumor. from misinterpreted data from the aerial monitoring site Flightradar24taken up by certain media and amplified by the AI.
Experts from Dartly emphasize that chatbots are strongly dependent on the media to verify the facts. but that they struggle to keep up with the rapid pace of news in time of global crisis.
They also alert on the distorting impact that these AI can have, while users rely more and more to learn.
Further reading: The United States released $ 30 million for the Gaza Humanitarian Foundation – Democrat primary in New York: here is the young left candidate who causes the surprise – Flippa-ball | When the grandmothers play the elbow – Valais banks school smartphones – Airlines reduce their service from the Middle East.
