Not just in the Navy, ships in the merchant marine are also “she”. Even small recreational boats are.
many languages Welsh/Cymraeg for example assign gender to inanimate objects and nouns - so in Welsh words like Afon (river) and many country names - “America” and “Awstralia” are feminine, whilst afal (apple) is masculine.
Your argumentation is correct but you are not blaming gendering of objects but the personification.
It’s slightly different than than using a gender, if I use she for a rope I’m not perceiving it as living.
But perceiving a thing as a someone is Indeed an issue.
Gendering is not a problem as long as we know that we refer to an object, in French the AI will be female because an application is female
Cette intelligence artificielle...
So if we applied your logic with gendering it was false because it doesn’t implies that the writer perceived the AI as “someone”
For the legal and moral issues around AI and automated system it will take time to resolve them but today it’s still the one who’s in charge of an automated system which is responsible for the damages it may cause.
For the case of cars we can imagine that if the AI is maintained by the constructor and if the owner haven’t done anything wrong with his care, Tesla will be responsible of accidents.
I’m not sure if they will even need to change laws
1 Like
I respectfully disagree. Referring e.g. a ship as female is personification. It is the very definition, grammatically speaking. Why is that wrong? Nobody thinks a ship is alive (not in a certain sense, anyway).
Chat bots even get names. That’s personification too.
I get that when, say, you make a phone call, or use a service desk on a website, that you may, due to this personification, not realise what you’re talking with, but let’s rather address that issue.
Perhaps one just needs some (more) regulation to get companies to make it clear if they’re using this tech. Then again, if it works and you get (correct) answers, does it really matter?
PS: I’m not ignoring some very pesky legal issues with AI, e.g. when it comes to creativity and prior art, or copyright, etc. Don’t be a strawman. My point above is about personification. #preemptingthebacklash
1 Like
From an English speaker point of view it’s the case, but we should not generalize to every languages.
That’s why I think that’s not the same, but sure how we speak about stuff reflect how we perceive it.
And yeah giving a name and a gender to a program is kinda the same as we do with stuffed animals and if the thing start to speak there’s a too little gap before thinking it’s alive.
We have all heard about LaMDA and the engineer who was persuaded the AI was conscious.
Machine learning replicates organic behaviour and it’s hard to inspect and understand a big neural network but people should easily keep in mind that we cannot ” hurt computer’s feeling ".
My personal concern would be for people with mental health disorders, they are at more risk to believe that the AI is a living being
1 Like
That is essentually what will be the main problem as we are entering a period where everybodies environment will be dense with AI.
People with mental health disorders problem is the mental health disorder and not AI. The artificial enviroment that we are heading to, could very well effect societies sanity as in effect as a human much too them is not real but portrays to be real.
Just like psychoanalysis which Freud was absolutely adament it should never be used for self advancement and now proliferates daily life from self help to propaganda its used to sell absolutely anything and everything.
AI is the ultimate interactive sales tool and from product to popular politics, missinformation and nasty rheteric its going to get heavy use and be pervasive for whover wishes to use or missuse and 90% of humanity is likely in a very strange position of being outclassed and outsmarted by something that as a human is not ‘real’
I wouldn’t have concern specifically with mental health disorders, but what it may do long term to the mental health of the 90% is something we are probably going to have to experience, to see.
The amount of chatgpt adverts in this thread masquerading as people is quite impressive
2 Likes
Yes, currently AI cannot give 100% correct answers, but this is due to the fact that there is no contract between artificial intelligence and HA, there is no standard that would allow them to interact and check each other, even config validation without rebooting was only completed recently. Python is not very suitable for this, because if the neural network writes crap, then the compiler will not check. Well, plus, the neural network must first start writing unit tests, for this it is necessary that the docks and standards of custom components be on a completely different level. For example, to have JSON schemas and UI HA would have even more separation between the HA server and the UI. You need to first transform UI HA into a design system, for example, begin with simple UI components. So that UI is built from the design of the system. Then a good additional training of the neural network is required so that it learns on these components
So what you basically is saying is that there is no such thing as AI
AI and home assistant will be work together as the best smart home in the world and I will be proving to you guys right to the end check a my post I keep everyone updated there.
By AI if you meant ChatGPT it’s not something we want to depend on, it might be an interesting integration, but we prefer open source stuff and things that can work offline.
Having a broken home automation just because something doesn’t work outside is quite a mess
Google Assistant is not open source and can not work offline. However, before you point out that OpenAI is young, it’s worth noting that Microsoft is actually older than Google. Additionally, there are open source alternatives available. It’s worth mentioning that Home Assistant was once one of the top trending projects on GitHub. If you check out the current trending projects, most of them are chatGPT or AI open source. By the way, if you’re looking for an auto GPT model, it’s likely based on ChatGPT
1 Like
Personally I’ll never use such arguments, anteriority and authority have no causality with quality.
And for Google Home it’s an integration for voice assistant, it might be a good way to integrate langage processing AI into HA.
HA Will not depends on it because voice is just a user interface, you can always fallback to graphical user interface.
But using a externalised logic for automation, scripts and scenes will not happen because those things need to stay local.
AI might help to analyse the way a user interact with it’s home and help finding triggers to create automations also.
Okay, if you have time, please do some research on this stuff. My guess is that you don’t know what plugins are, which most people have nicknamed the OpenAI App Store. I’m also guessing that you don’t think it can talk like Google. However, it has been talked about on speakers and phones. It just doesn’t have a device yet, which is good because then we can put it on our own devices and not rely on Google Home or their products. I have a post on this forum about the technology growing over last 6 months. 
If your not an believe then don’t post on there please.
This can work On your servers
1 Like
I guess you just misunderstood, I was just saying that the whole point of HA is to not rely on the cloud.
It can work with it but mustn’t depends on it, your home automation need to be reliable, voice assistant, chatGPT, nest products etc need internet to work, moreover they need the servers and the cloud services to be updated.
If you base all you home automation on those you will have no control of you devices if there’s an internet issue.
We don’t really care if a voice assistant stop answering, it’s not an issue either if data science tools stop to work.
But the automation part need to remain static and local (even if AI generated) to keep reliability.
I never said that it will not be integrated neither I said anything about the model itself, just that HA will always need to be able to work offline and thus cannot depends on cloud services
The issue with self hosting is the computing power it needs but but onboarding AI models into IoT stuff is something plausible since it’s already done for things like LPR, Microsoft kinect and phone’s facial recognition.
I guess that in the future language model will also be able to run on IoT stuff, even tho the learning will be processed in datacenters the finished model will run offline.
There’s couples of issues to achieve that (like ram needs) but in the future they might be solved.
- Will AGI be a part of home assistants very soon, given the progress that OpenAI has achieved in the past six months with its Chat GPT, as well as the new ‘OpenAI plug-in store’ (similar to Google Assistant apps) and a new project called ‘Auto GPT’ that was announced this week?
- Yes
- No
- Maybe
0 voters
Guys, this thread is about using AI to answer forum questions. And the current answer is: No you cannot use AI to answer questions.
Anything else regarding AI is purely speculation.
1 Like
I just don’t understand the fears of artificial intelligence in this topic, because in the future artificial intelligence will be able to develop program code for the Home Assistant, this is not some kind of task that listens to the user and answers questions. It can be some configs, a new beautiful UI, the generation of beautiful pictures for the style of a particular house. All this can be done through the usual code review of scripts that artificial intelligence can write. You can also make a sandbox mode. It is not very clear why many people trust custom plugins without even going to the github page and without reading the reviews, but at the same time they will not trust the code written by artificial intelligence. But there is currently no such code. Now is better to think about new lever of HA core isolation, and, for example make UI ready for AI auto edit, and not giving AI rights to reboot HA for example, and install plugins
2 Likes