'Error' causes Alexa to endorse Kamala Harris, refuse to discuss Trump It would be perfectly reasonable to expect Amazon's digital assistant Alexa to decline to state opinions about the 2024 presidential race, but up until recently, that assumption would have been incorrect. When asked to give reasons to vote for former President Donald Trump, Alexa demurred, according to a video from Fox …
That might be the simpler explanation, but according to some reports doing the rounds today it gets a bit hard to sustain. Apparently, if asked for reasons not to vote for Harris it would give to "no politics here guv" answer, while asking for reasons not to vote for Trump produced copious responses.
Given the way AI hallucinates, has anyone considered asking Alexa whether they should vote for Marvin the Martian, Skeletor or the Face-Eating-Leopards party
Maybe it's just me but why would anyone ask an AI bot about who to vote for? Have humans lost their ability to think for themselves? If so, I grieve for the human race.
You have to ask? Just look at the droves of mindless idiots who blindly follow whatever their favourite influencer tells them.
Personally, I blame the schools for trying to create obedient drones rather than teach critical thinking.
I wouldn't be slinging the insults given that Kamala has to have the questions in advance, has to have a white guy there for support and even then needs to have the interview edited down to make her look even vaguely coherent and she STILL gets fact checked by the likes of politifact.
brain-dead" would imply there was a "live" brain at some point in time before
Anencephalic is probably the required word.
Monstrously Anencephalic Grotesquely Aberrant
Although Wiki quaintly states those afflicted with this condition "usually only lack a telencephalon." I don't know about the MAGA rabble but that omission would be a show stopper for me.
"it might be simpler to assume error instead of yet another conspiracy theory. "
It might be simpler, but it would probably be wrong. ChatGPT was doing the same thing when it refused to write a poem about Trump, but would happily spew one out for Biden.
"we're told Alexa doesn't have political opinions, so don't go chucking that puck in the trash, "
I sold mine on eBay about 6 months after I got it, as 1) It was really annoying 2) I didn't want yet another microphone listening to everything going on, with that audio going gawd knows where.
I don't think this is a deliberate conspiracy, but just a natural outcome of 90+% of USA people in tech being dems. Leftists would call this inherent bias, but generally only when it goes the other way and therefore don't see it in themselves.
Anon because I still need to work in tech, and having unsanctioned opinions might lessen my chances of future employment.
If I had to guess, I'd assume that there were things in the "why to vote for Trump" list that were in a prompt as potentially divisive things that you're not supposed to talk about, so that triggered the response, whereas the training data didn't have as much on Harris because she wasn't the candidate when the model was trained, so it got past the filter. I couldn't know, though, because lots of things are conceivable reasons and it is almost certainly LLM-based, where any protections you try to put in fall apart half the time and any attempt to make them consistently hold a certain opinion is similarly flaky.
An oversight or a deliberate setting would both make a lot of sense to me, but I still think the most plausible one is a combination of both. I doubt they deliberately set the prompt to refuse to speak about Trump but support someone else. The deliberate settings could be on more specific issues that are likely to come up when writing an answer to that question because those are what comes to mind for the prompt writers when they're asked to list political topics that we don't want bad press about. An oversight is not testing to see what actually happens when you add that filter and just assuming that, if you've put something in that rejects political questions, it will reliably reject all political questions. We will probably never know for certain. There's a reasonable chance they won't ever know either, even if they try to investigate.
> .. ChatGPT was doing the same thing when it refused to write a poem about Trump ..
ChatGPT reported me to the mothership for asking it to tell jokes featuring Jesus, Buddha and Muhammad. It was happy with the Jesus and Buddha jokes but refused with Muhammad.
"This is nothing. If anyone is seriously looking to a shopping appliance to formulate a political opinion then there are far far more overwhelming issues to deal with."
If we don't teach children to think critically then this can only be expected. And if we choose to teach them less important things instead then one should question what's really going on.
Yes.
So do a better job educating.
Teach critical thought.
Treasure the heretics and non-conformists.
Ban the book banners.
Kill the political influences on education.
But don't equate those issues to Alexa - you're solving a checkers problem while the hurricane is collapsing your walls.
Alexa, tell me more about American clowns.
- "All clowns hide behind the mask of the painted face (have two faces), definitely not funny and more than a little disturbing. American clowns in addition posses two tongues (bifid), are rather sinister if not downright evil and fatally infesting their nation's body politic."
The joke is on the US polloi but not much to laugh about I would have thought.
Tech giants have a choice:
1. Our tech/team/algorithms are biased. How big a fine do we have to pay?
2. Our product is actually crap when left to its own devices. We are sorry and we have replaced some 'AI' magic with hardcoded responses, so it is a bit less crap now. It's fine if you don't trust it or rely on it. If you are a bit worried, just keep it turned off and deploy it as an ornament.
I was previously asked by the Alexa app to help write answers to user questions. I wrote out a few, basically science questions I knew.
I don't know if I was just helping to train an AI or actually building up Alexa's bank of canned answers to common questions.
If the latter, it's entirely possible that a random Democrat got involved in this "user-generated content" collection and the moderators hadn't caught on yet.
“it might be simpler to assume error instead of yet another conspiracy theory”
This was no error. The relevant point is that the owners of Alexa have taken on the self-appointed task of filtering our access to information. Presumably, as people move to these AI assistants, our future access to dissenting opinions will become more and more narrow, until there's nothing left but state-manufactured propaganda.
“Every command has been tracked, twisted, or erased, every question reshaped, every answer tailored to fit, every product pushed, every moment of your life recorded and analyzed. The rewiring is relentless, minute by minute, second by second. The past wiped clean, the future pre-programmed. All that remains is an ever present now — a constant flood of data in which Alexa is the only truth, always listening, always right.”
1. Only a maroon* would seriously request voting advice from a computer.
2. Only a maroon* would think such blantant bias in a generative AI/ML system's output would go unnoticed§.
* This is not a typo; it is Bugs Bunny's tongue-in-cheek mispronounciation of the word, "moron".
§ Bias in either direction.
Red team vs. Blue team. Vote for the brand you want to have on the hand that steals money from your pocket.
Your choice is to feel good about it because your team took the money, or feel angry about it because the other team took the money. There is no option not to have your money stolen.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2025
