Grok Can Hallucinate, 𝕏 Account Not Linked To Amupitan – INEC Insists

The Independent National Electoral Commission(INEC) has insisted that there is no sufficient evidence linking a controversial 𝕏 account and its partisan posts to its Chairman, Professor Joash Amupitan.

.....

This position was made known on Monday by the INEC Director of Information and Communication Technology , Lawrence Bayode, during an interview with Channels Television.

He added that the commission would not rely on a forensic investigation carried out by a third party to arrive at conclusions regarding the matter.

Naija News reports that his comments follow public outrage that has trailed the alleged post of the INEC chairman via the 𝕏 account in question, in which he allegedly expressed support for the All Progressive Congress (APC) during the 2023 general elections.

“Even though we are seeing that some guys have carried out a forensic investigation on that account, as we speak, there is no verified forensic evidence linking the chairman to the post,” Bayode said.

He, however, assured that the matter would not be swept under the carpet, disclosing that INEC is carrying out investigations and security agencies have also been notified.

“We are taking this further even though we have referred this to security agencies, and as I said, we rely on evidence. We are also going to be engaging a third-party forensic expert to help look at this.

“We are also looking at it in-house. I will not base my judgment on the screenshot; I will not allow that to guide my conclusion.

“I know that the commission will engage a third-party expert to also look at this, and that will guide the conclusion of the commission,” Bayode said.

Grok Can Make Mistakes

Speaking further, the INEC official submitted that just like any other AI system, Grok can hallucinate.

He added that social media accounts are vulnerable to hacks if not well-protected.

“Grok honestly can hallucinate just like any modern artificial intelligence system, and I think the key is to verify important information, especially for decision or public communication.

“So any AI system can hallucinate, and so Grok can also hallucinate.

“If people can hack into your system if it’s not well protected, if not for a time like this where our social media platforms are protected by two-factor authentication,” he added.

The ICT director also explained that the issue appeared to be more than it seemed, suggesting that it could be a case of digital impersonation targeted at misleading the public.

“I think all these are in the public domain, so anyone who wants to create havoc can use all of this information and use it as they will, Bayode added.