AI Hype vs Reality
This week: AI hype vs reality, Meta continues to prioritise profit over people, and why it's not just good to talk - It could also save you money.
3 - AI hype vs reality
“Ironically, one of the biggest AI-related security issues is the obsession with fictional threats while neglecting fundamental security best practices.”
Source: Socket (and shared by Greg Schaffer)
What?
MIT recently published a report claiming 80% of ransomware is AI-driven, but it has been widely criticised by security experts for lacking evidence and overstating AI’s role.
The report was sponsored and co-authored by a cyber security company that sells a tool to defend against AI attacks.
Critics say that the research measured survey participants perceptions of the risks rather than hard evidence of actual attacks, which misleads executives and distorts priorities.
MIT has now taken the report offline.
So What?
We can’t let AI hype distract us from the real risks that we face today.
While AI gets the headlines, we must remain focused on the ‘old reliables’.
And the oldest reliable? People being fooled by scam emails and SMS messages.
2 - Meta continues to prioritise profit over people
“10% of Meta’s revenue apparently comes from fraudulent adverts and illegal content”
Source: TechRadar
What?
Leaked documents suggest Meta earns up to 10% of its revenue (around $16 billion) from scam ads and banned goods. Despite internal warnings, many fraudulent ads remain active until flagged with near certainty.
Apparently, even when Meta is 94% certain that an ad is a scam, it will not remove it from its platforms. Instead, it will increase the fee that it charges to the scammer to keep the ad online!
The documents also suggest that Facebook weighs the value of the income that it earns from these scam ads against the likely cost of the regulatory fines that it could receive by not removing them.
Capitalism at its best!
So what?
Meta executives (i.e. the people behind Facebook, Instagram, and WhatsApp) continue to prioritise profit over people.
And yet regulators and lawmakers in the EU are under increasing pressure to reduce the regulatory ‘burden’ on these companies.
These regulations may be a burden on large, profitable businesses.
But they are also (imperfect) attempts to protect us mere mortals.
1 - It’s good to talk. And it might also save you €100k.
“[The Irish University] paid the full invoiced amount to the bank account indicated in the fraudulent email, without checking independently with the supplier”.
Source: RTE News
What?
An Irish university was fooled into sending €98,500 to a fraudster after acting on a fake email that asked it to change a supplier’s bank details. The university failed to verify the change, breaching its own controls.
It has since enhanced internal procedures and retrained finance staff.
So what?
It’s good to talk.
Before setting up or changing the payment details of any supplier, you must phone the supplier on a number that you already have on file for them, and ask them to confirm their details.
It’s a simple step and an effective defence against the most common cyber crime.
This week’s roundup is a good reminder to separate the marketing buzz from what AI systems can actually deliver today. The story about Meta prioritising profit over people highlights how platform incentives can skew the development of responsible AI wthout meaningful oversight. I thought the note about the value of simply talking to others and building community around digital wellbeing was refreshing; the social angle often gets lost amid technical breakthroughs. It's also wise to remember that some tools are still experimental and may not save us money or time yet, but staying informed helps us set realistic expectations. Appreciating these nuanced takes helps ground the conversation wih a sense of pragmatism.