Yesterday morning Kenyans woke up to an apology by Safaricom, an apology majority refused to accept. The apology was occasioned by a series of tweets that indicated that one of the Safaricom’s social media customer care representatives had mistakenly used the official Safaricom Twitter Account @SafaricomPLC and @Safaricom_Care to send personal tweets, including a tweet voicing an opinion over the successor of Bob Collymore who recently passed on. According to the apology by Safaricom, the tweets were due to a system glitch that had been rectified by the time they were sending that apology.
We apologise for the series of tweets that went out from our official accounts last night.
We had a system glitch that has since been rectified.
— Safaricom PLC (@SafaricomPLC) July 4, 2019
Kenyans rejected the explanation that it was a system glitch (although those who bought the explanation attributed the tweets to Zuri – Safaricom’s AI customer care representative) whereas majority sarcastically accepted the explanation by giving a thumbs up to the glitch for tweeting the bitter truth.
But what was tweeted was bitter truth. I’m a patriotic Kenyan but our own Kenyans have disappointed us with the culture of looting from companies instead of building them.
— KISIKA SAM (@KISIKASAM) July 4, 2019
UPDATE: Safaricom has since clarified how the System Glitch tweeted the personal tweets. Check out the article Safaricom clarifies about the Safaricom System Glitch that affected their Twitter Accounts for details.
After thinking through these two possibilities namely that a customer care representative had forgotten to logout of Safaricom Twitter Account before sending the personal tweets, and that there was a system glitch, I can’t help but come up with Safaricom Twitter Account was hacked as the most logical explanation. But first let’s examine the demerits of each of the two explanations put forth so far.
A customer care representative sending personal views via Safaricom Twitter Account is more likely than Safaricom’s explanation of a system glitch, as the tweets were obviously those of a Kenyan complaining of several things, including asking Safaricom to help with his/her home Internet that was not working.
The problem with accepting that the person sending the tweets was a Safaricom employee as social media customer care representative, or any other technical/managerial member who had access to the Twitter account, is that the said person could send several tweets mostly asking the same account to help solve a problem on one of the Safaricom services, without noticing for hours, unless the said customer care representative was begging to be fired. What normally happens, and this I say from personal experience as I also manage our Twitter Accounts, is that when you realise that you have sent a tweet using a wrong account, you delete the tweet immediately. Most cases I am able to delete the tweet before anyone on Twitter has noticed it to screenshot it. But when I go through the timestamps of the tweets that were sent via the Safaricom Twitter Account, I notice that the tweets were spaced minutes apart. For example the above tweet was sent at 23:39 of July 3rd 2019, whereas the one on the next screenshot was sent at 23:40 of July 3rd 2019.
It is thus very obvious that the person who was sending those tweets was not only doing what he/she knew, but was doing it in such a way as to be noticed. So maybe Safaricom’s explanation of a system glitch is afterall the true position.
A system glitch implies that the system was tweeting by itself, which would be an outrageous claim if made at the dawn of Internet and social media. Today however, when we have advanced AI that can hold conversations with humans without the human being able to detect that he/she is conversing with an AI, such an explanation is perfectly acceptable. Actually we have had the bots become so human that on several occasions they have been turned off for being racists. In that light, it is feasible that Safaricom’s chatbot, Zuri, that is known to offer help to Safaricom subscribers over at Facebook, had a glitch and decided to send out a series of tweets, having learnt of what to tweet from interacting with Safaricom customers in the period it has been in operation. Although the argument of a glitch is a possibility, I highly doubt that Zuri is able to be as witty as the person who was sending the tweets. Consider this tweet that was sent as a reply to Dennis Itumbi for instance;
I refuse to believe that AI has become that witty and sarcastic, with a dark sense of humor, and that’s the reason I only buy the explanation that Safaricom Twitter Account was hacked as fronted by @Zakariasharka1,
The kcse class of 2015(mass exam cheat) are responsible for all this hacks.. we know its them on attachment (simple finger count) ??
— Zak.enya??(??) (@Zakariasharka1) July 4, 2019
It is just last week that Safariocm again explained that due to a System glitch, members of the public were able to buy data bundles and minutes worth close to a billion shillings. With another hack now focusing on the organization’s social media accounts, one cannot help but wonder, “how safe is our money on MPESA?”