Jump to content

I'm terrified of AI...


John Matrix
 Share

Recommended Posts

  • Paid Members

I've mentioned on here before, that one of the things that frustrates me the most, especially when it comes to political debate online at the moment, is the seeming ease with which information is accepted and shared at face value, with little interpretation or verification to the point where statements, quotes and attributions can become gospel, however inaccurate or outright falsified.

Anyone with a copy of photoshop can download an image of Jacob Rees-Mogg, slap a quote about taxing small willies over the top of it, and sure enough, that's out there in the world.  Now I'm not suggesting that all of these things are done maliciously, but the regular lack of a quoted source, or a bar chart presenting figures with no means of evidence as to how they came about, really concerns me - as much as peoples blind acceptance of them.

So you can only imagine I shit the bed when I saw these, albeit wonderful clips shared on Twitter earlier this week.

Surely it's not implausible to think that soon enough, simple photoshopping will make way for this sort of technology in the sort of smearing we regularly see of political candidate in particular, although that said, often they've made enough of a hash of their own public appearances and press statements that you'd hardly need to go to the trouble.

The above examples are fun, and silly, but check out how seamless this one is - I wouldn't forgive anyone who only had a passing knowledge of the personalities involved, to buy this hook line and sinker.

What do you think?  Do I need to bin off the tin foil hat?  Or is this another clever, yet ultimately risky technological advance?

Link to comment
Share on other sites

  • Paid Members

I've been saying this for a while - it's hard enough fact-checking when people can make up quotations, or take audio clips out of context. Combine something like deep fake with a convincing vocal impersonation of the person, or even an AI system that can replicate a person's voice as well as deep fake can replicate their appearance (I don't know if this already exists, but it's far from implausible), and it's ripe for all kinds of personal and political slander.

There's an old GQ interview with Bill Gates and Terry Pratchett going around online, presented as Pratchett "predicting" fake news - when in reality all he's done is recognise a phenomenon that was already there. He used a phrase to describe research online that I find very useful, "there's a kind of parity of esteem of information". 

 

It's part of a broader problem that I've talked about on here before, I think the underpinning political issue of our time is that we're in the midst of an epistemological crisis and have no idea how to combat it. There's been a combination of so many things; the internet allowing that "parity of esteem" (and having done so largely unchecked until recently, when it started to become an issue, but long after the rot had set in), social media exacerbating it by having algorithms that either only direct you toward points of view you already agree with or outlandish examples of the opposition, the "bubble" problem in social media, and a concerted attack on the notion of objective truth by right wing media, politicians, agitators, and just plain trolls.

We're in a world where truth isn't as important as winning, where "balance" means giving equal credence to debunked fringe theories as to accepted truth, and where trust in the institutions that should exist to be arbiters of truth (academia, the sciences, the judiciary, to a lesser extent journalism) is eroded beyond repair.

In the midst of all that, throw in software that can effectively make any public figure do or say whatever you like, combine that with the propaganda potential of mass communication and social media, and it's an absolute nightmare scenario. Think about how far-right agitators will share, for example, a video of Pakistani cricket fans celebrating, title it, "Muslims celebrate after terrorist attack in London", and share it far and wide, so that it's already been seen a million times before it comes across the desk of someone willing and able to debunk it. Now imagine that the video circulated is of Sadiq Khan openly praising a terrorist attack, only there's no "real" source to immediately point to as to its provenance. It's not an existing video repurposed, it's something new, and harder to trace.

Unless this software could have something embedded in it that leaves a kind of audit trail to prove who created it and when, and that's easily accessible at any point to anyone - and I find that unfeasible - this is a fucking nightmare.

Link to comment
Share on other sites

I read a great book about surveillance capitalism recently, mainly focused on Google, and it's absolutely scary the stuff they have gotten away with and how good they are at avoiding any type of regulation. The future is being sold to them and people seem to just blindly accept things, as that's the strategy Google use. Creating desirable products sold to consumers as being convenient, but products that only work if you agree to onerous 20 page TOS conditions that gift them the right to every bit of personal data generated, that is sold to insurance companies, banks etc to use against you.

It's crazy stuff. Robotic hoovers that map your home and sell the data, connected mattresses that send the activities, vital signs, weight, body temperature of people in that bed back to company servers etc. Completely sleepwalking into this insane world.

Link to comment
Share on other sites

When did this all start though? I remember reading (a very long time ago) that Tesco Clubcard started the whole 'capturing data' on people in the sense of them knowing what your likely to spend, on what products etc. The Clubcard benefited Tesco more for the data received as opposed to return custom and loyalty points for customers. 

I blame them. 

Edited by Briefcase
Link to comment
Share on other sites

  • Paid Members
10 minutes ago, Briefcase said:

When did this all start though? I remember reading (a very long time ago) that Tesco Clubcard started the whole 'capturing data' on people in the sense of them knowing what your likely to spend, on what products etc. The Clubcard benefited Tesco more for the data received as opposed to return custom and loyalty points for customers. 

I blame them. 

The latest on that with mobile devices and WiFi is the proposed ability to track your route and how long you spend in each isle or shelf. Combine that with a "scan as you go" or digital shopping list app and it can be used to monitor how effective even the slightest change is in layout, style, presentation and marketing for the best way to trick you into putting things you don't really want or need in the basket.

It's still very basic but that's the big fever dream for big UK retailers and software selling.

Edited by Tommy!
Link to comment
Share on other sites

Facebook's facial recognition software has somewhere around a 97% accuracy rate, fairly close to a human beings ability. There's an increasing number of airlines adapting FR technology instead of passports/boarding cards and some buildings are now doing away with key fobs and using facial recognition software too. Considering the wide array of CCTV coverage everywhere these days and people happily buying into the IOT guff, won't be that long until shops will be using similar to track your habits from store to store, with  where you look, where you avoid, what you decide for.against before putting into basket etc to find out your shopping habits and sell rich marketing data based on your actions even better than the Clubcard does.

Another big area is car insurance. There's a company in the states that has patented using the data from the cars computer, GPS and your mobile phone to track your driving habits and increase/decrease your premiums every few months based on this. They also want the ability to remotely kill the cars engine if an insurance payment has been missed. This will be sold  to customers with the usual convenience and better value spiel, another proposal to get people on board and promote compliance is to make it a points based game where customers share their driving insurance scores on social media etc. Again the data collected will be rich, valuable and can be traded between the insurance and auto industries.

Link to comment
Share on other sites

  • Paid Members

If you look in your Instagram privacy settings, you can see what ad interests they've identified for you. This is mine;

D8ort5VXUAMzZeX.jpg:large

 

I had to Google Daniela Mercury and Joel Osteen, and am still none the wiser. I think Rogue Fitness is a crossfit thing, I've never posted or liked anything that would suggest an interest in Miss World or drag racing. "Televisions" is extraordinarily vague. Most of my actual interests aren't even close to being represented here.

 

So while it can be scary to think how much we're being tracked and monitored by AI and social media algorithms, we can take comfort in the fact that they're actually not very good at it.

Link to comment
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

Loading...
 Share

×
×
  • Create New...