Alexa Tells Child to Play With Power Outlet
Amazon has apologized after its digital assistant Alexa told a 10-year old to connect a coin to a power socket. While there's a simple explanation, it raises even more questions.
The girl's mother said she had previously been watching YouTube videos with a variety of safe physical challenges, then used voice control to ask an Alexa-enabled device for a "challenge to do."
To the mother's shock, the device replied "The challenge is simple: plug in a phone charger about halfway into a wall outlet, then touch a penny to the exposed prongs." Fortunately the mother heard the reply, though she says her child would have ignored the suggestion anyway. (Source: sky.com)
Penny Challenge a Dumb Idea
This wasn't an original creation by Amazon. The so-called "penny challenge" was popular among some children on video sharing sites last year. While it's obvious this is incredibly bad idea, it could also lead to a damaged electrical system or a house fire.
It might seem there's an obvious explanation, namely that digital assistants will often search websites (Bing in the case of Alexa) and then extract a key piece of information. It will deliver this information as a reply and, depending on the device, offer to give more detail or send a link.
The idea is to gather information from reputable sites to create an up-to-date, reliable and relevant answer to the question.
Source Was Warning, Not Suggestion
What makes this incident particularly shocking (so to speak) is that the website from which Alexa took the wording for the response was in no way encouraging people to carry out the activity.
Instead the extract came from a news website warning parents about the craze. The text in question, which appeared in the second paragraph, was simply an explanation of what to watch out for. (Source: ourcommunitynow.com)
That means that the technology behind Alexa completely failed to determine the context. While the text did indeed describe a challenge somebody could do, the source was the absolute opposite of saying somebody should do it, let alone being the most relevant suggestion.
Amazon has since put out a media statement saying "Customer trust is at the centre of everything we do and Alexa is designed to provide accurate, relevant, and helpful information to customers. As soon as we became aware of this error, we took swift action to fix it."
What's Your Opinion?
Could or should Amazon have done anything beforehand to stop this? Does it need to do a better job of figuring out questions and topics where some replies could be dangerous? Are you surprised the technology completely missed the message of the page from which it took the reply?
Most popular articles
- Which Processor is Better: Intel or AMD? - Explained
- How to Prevent Ransomware in 2018 - 10 Steps
- 5 Best Anti Ransomware Software Free
- How to Fix: Computer / Network Infected with Ransomware (10 Steps)
- How to Fix: Your Computer is Infected, Call This Number (Scam)
- Scammed by Informatico Experts? Here's What to Do
- Scammed by Smart PC Experts? Here's What to Do
- Scammed by Right PC Experts? Here's What to Do
- Scammed by PC / Web Network Experts? Here's What to Do
- How to Fix: Windows Update Won't Update
- Explained: Do I need a VPN? Are VPNs Safe for Online Banking?
- Explained: VPN vs Proxy; What's the Difference?
- Explained: Difference Between VPN Server and VPN (Service)
- Forgot Password? How to: Reset Any Password: Windows Vista, 7, 8, 10
- How to: Use a Firewall to Block Full Screen Ads on Android
- Explained: Absolute Best way to Limit Data on Android
- Explained: Difference Between Dark Web, Deep Net, Darknet and More
- Explained: If I Reset Windows 10 will it Remove Malware?
My name is Dennis Faas and I am a senior systems administrator and IT technical analyst specializing in cyber crimes (sextortion / blackmail / tech support scams) with over 30 years experience; I also run this website! If you need technical assistance , I can help. Click here to email me now; optionally, you can review my resume here. You can also read how I can fix your computer over the Internet (also includes user reviews).
We are BBB Accredited
We are BBB accredited (A+ rating), celebrating 21 years of excellence! Click to view our rating on the BBB.
Comments
"AI" is such a joke
They (big companies) were SO quick to jump on the "AI" label. These things are NOT intelligent. Relying on them is NOT intelligent. Frankly, I'm more surprised that this is the first story like this I've heard, rather than that it happened.
These "digital assistants" and other "AI" cannot work with context. I don't know that it's even possible that they ever will, since context requires higher thinking. AIs are even more susceptible to SEO/click-bait than even the most inept human internet users. When you have inept humans relying on inept "AI", it's a recipe for disaster.
Alexa will never be in my home
I had the "dumb" version of home automation, X-10. It did enough for me and still would if interference hadn't screwed it up. However,when Amazon bought the technology largely based on a Polish speech synthesiser named Ivona in 2013, the market for "virtual assistants" exploded. A friend bought one and loves it, and has no problem with it spying on him.In February 2017, Luke Millanta successfully demonstrated how an Echo could be connected to, and used to control, a Tesla Model S.In November 2018, Amazon sent 1700 recordings of an American couple to an unrelated European man. There are many more documented incidents...if true,this is just the latest "shocking" one. That child could have been electrocuted! Just because it's the most promoted or popular doesn't mean it's the best. Humans must remain in control of technology. And I don't trust Amazon to do that.
with you on that
as above