TWIT 922: AI Hustlers

Beep boop - this is a robot. A new show has been posted to TWiT…

What are your thoughts about today’s show? We’d love to hear from you!

I am guessing someone from Google prodded Dopey Joe and told him AI was evil. I’m beginning to think it will only be seen as good for the world when Apple release their product.

1 Like

I think it will only be seen as good, when it starts obeying current laws.

The reason it has been banned in Italy is that citizens found their data was being handed out by ChatGPT, even though they had never given OpenAI permission to use their data, which is a GDPR offence, which could bring big fines with it, if the citizens and the Data Protection Commissioner in Italy are successful, it would mean that it is also illegal in the EU, the UK, Canada, Iceland, Japan and other countries with similar levels of data protection.

Open AI knew about the law when they started training their AIs, the laws have been in effect since before the company existed, yet they decided to ignore the law in order to make a shortcut to market.

This has been seen again and again by Big Tech and Silicon Valley startups, and has nothing to do with AI per se, just Tech’s disregard for the law, until the lawyers and fines cost more than obeying the law - and they then often claim that it is too late to change and offer risible alternatives, a la YouTube.

If AI can stay within the existing laws, and it can provide good results, as opposed to presenting bad results with confidence, I don’t see any problems with so-called AI, as it is at the moment. But that brings a willingness of the owners of the technology to do the right thing from the start.

1 Like

Humanity’s downfall will be the result of its own intelligence.

Just because you can do something, doesn’t mean you should.

I wonder if this will kill me?

F*$% around and find out.

1 Like

I don’t thing government should have the power to say what you can and can’t share, you should have the option to opt in if you wan

Once the hype cycle has died down and we start seeing less parlor tricks and more actual uses for AI then you’ll see it just become part of everyday life not really something what “accepts”. People don’t ‘accept’ auto-correct they just use it without thinking. Things like that are the future of AI, things that nobody thinks about but they use every day.

1 Like

We will cease to be human thinking machines and rely on AI to do all the thinking for us (already happening). Imagine where this will lead us in 100+years.

Reduced to human consuming machines being fed by AI that does all the thinking for us.

Cynical, or the reality that is already in front of us?

1 Like

Are you being intentionally unthinking here? Of course you can opt in to something if you want, but a profit making business has no limits on what it will do without the government setting limits on it. The GDPR was designed to take the power to collect personal data away from powerful monopoly like businesses and place it back in the hands of the people who they extract the data from without permission. The best result might be that the company makes clear what your benefits are for co-operation… but those benefits rarely exist and most people wouldn’t play the game for the few benefits they might otherwise gain.

3 Likes

You can share what you want, as long as it is yours to share. That is what the government have defined, within limits, no underage sex and other things shunned by society, in Germany, it also covers Holocaust denial and the glorification of National Socialism, for example.

So, it by definition, also says what companies can’t share, and that is your personal information, without your explicit authorization - they can’t use it themselves for purposes other than those they stated when they collected it and they cannot pass it on to others, for free or for money, without your express permission.

It is your data, you decide who can have it and what they can do with it.

They are some exceptions, such as law enforcement being able to store information on you, if you are involved in a crime and the storage of that information for a set period of time after your conviction and release. Companies that you do business with need to be able to write you an invoice, so they need your information and must store your name and addres against all invoices sent to you, for tax purposes. But you can request that any other information be deleted, if it is no longer relevant - you have ceased your business relationship with them, for example. The information stored for tax purposes, however, must be destroyed after 10 years (Germany, I assume the UK has a similar period of time, after which the information must be destroyed).

The basis for information ownership in the EU (and the UK, at the moment) is that personal information belongs to the identified person and, within those certain limits such as those mentioned above, it is their’s to do with and share as they please.

If the bank has your transaction information, they cannot share it with or sell it to third parties. If a telephone company or ISP has information about who you call or which sites you visit, it is illegal for them to sell that information, they cannot share it with or sell it to a third party. They also cannot store it outside the EU/UK, unless the destination land has an equivalent level of data protection - which currently excludes the USA, because they have managed to break 2 separate agreements, so far, on the storage of that information, thus making it illegal for a US concern to hold information about European citizens.

They used to be able to hold it within Europe, but the CLOUD Act nixxed that and now it is illegal for US companies to hold any data over EU citizens - which is why Facebook and WhatsApp are essentially illegal in the EU and the UK; you can give them information about yourself, but not third parties, without their express permission, but Facebook never bothers itself with ensuring that information has been obtained (WhatsApp loads your entire address book from your device, when you use its service, even though they have failed to obtain the permission to upload that information from everyone in your address book first. Others, like Signal, AFAIK, store a hashed version of your phone number and your contacts phone numbers, when there is a match you are informed, they don’t store the names of your contacts until they have signed up and given Signal their names freely. WhatsApp does have a business app, where it doesn’t, allegedly, pillage your device for information, but very few people use it, so WhatsApp is quasi illegal in Europe and you could be in trouble if someone doesn’t want Meta to have their data and they find out you gave it to them without first obtaining that person’s permission).

Again, this has nothing to do with AI per se, just the way Big Tech and Silicon Valley ignore the law.

Edit: @PHolder said it much more concisely than I did. And this applies to the EU, UK, Japan, Iceland and a few other countries around the world that have similar levels of data protection to the EU.

2 Likes

That is certainly one possible future, if we stop critical thinking. But, as long as pupils are taught critical thinking in school… Oh, wait… :man_facepalming:

1 Like

I think there could be some kind of encoded file you keep on your devices that should be available to enhance your publicly shared information

Paul Smith-Keitley
Adobe Creative Educator

Maybe, but that doesn#t solve the problem of companies like OpenAI grabbing PII without permission and giving it out illegally to third parties.

If you have given permission is one thing, but millions haven’t, yet OpenAI still pulled in that PII.

Again my question would be where did they get it from? If it was publicly accessible I would say no foul. Just playing devil’s advocate here.

Paul Smith-Keitley
Adobe Creative Educator

They got some of it through idiots entering it into ChatGPT.

There have been cases of managers putting PII of underlings into ChatGPT and asking it to provide a summary. ChatGPT doesn’t forget what it has learnt…

Similarly, companies have found confidential internal information is being handed out by ChatGPT, because some idiot entered the company finance report into ChatGPT and asked it to provide a PowerPoint slide deck with a summary of the information.

Samsung was also caught, confidential information about its semicondutor production was found in ChatGPT.

I think it was Amazon or Google that banned ChatGPT for its programmers, because they were pumping proprietary code into ChatGPT and asking it to document or optimize the code. The lawyers jumped very quickly on that one.

They have to somehow come up with safeguards that confidential information is treated confidentially, or people can’t upload it in the first place. The same for PII. If ChatGPT recognises PII, it should stop and ask if the person has given their permission for the informaiton to be uploaded and also how widely spread information that is uploaded could be spread.

Given the examples above, I’m guessing a lot of people don’t realise that all the information they pump into these AIs is then freely available to everybody around the world.

If this is a reference to the President, could we please avoid this kind of discourse here. I’m not a fan of the previous president, but I don’t go around throwing insulting nick-names. Your argument should be able to stand on its own merits without resorting to this kind of stuff.

2 Likes

I agree, sorry, we should keep politics out of tech. FYI I am a foreigner so I maybe didn’t give the office the respect it may deserve. We are a bit less in awe of our politicians over here. Sorry if it offended you in any way. Many people over here are somewhat annoyed at his seeming interference in the Northern Ireland issues, it is after all nothing to do with him he is an American.

2 Likes

Thank you. No offense was taken as such - but it’s too easy for conversations to devolve into politics, especially given some of the topics that are discussed on Twit and elsewhere. I appreciate your perspective on this.

2 Likes

Not sure what Doc Rock was thinking claiming Alphabet to be the world’s biggest software company, when it’s market cap is amost $1T less than Microsoft.

Come to think of it, the presenters without Leo seem to be of the opinion that Apple and Google are the only game in town.

It seems to be US thing, where people believe Macbook to be worth the extra, in the UK they don’t have the same cachet, many tech people (non journalists, people actually working in tech) find them quite crazy priced and almost bling rather than real tools. Chromebooks are viewed as poor people devices