Thanks John. The article explains what I thought but couldn’t express. The joke on taking things literally was good.
ChatGPT is a highly evolved version of Eliza which – in some versions I saw – was under 100k bytes but still managed to surprise a few people.
Neither had a clue what it was talking about.
The big problem with detection is when a post is brief; 1 or 2 lines. Fortunately in our case it is readily detectable by anyone who actually knows about aviation, but it would be a hell of a problem for anyone running a forum for revenue (which is 90-99% of social media) since these will just fill up with it. One ski forum on FB has just been flooded with it…
What should one do to get 0% probability to be mistaken for a chatbot?
Ok the last line was partly referring to Peter’s “one-line posts are hard to identify”, but anyway the point will not too long in the future definitely be interesting to find out. Because if 95% is produced by a chat bot it’s harder every day to find out what is real
Clipperstorch wrote:
Brace for a new level of scamming.
not only scamming, but fake news “read” out on live TV by people who never knew what they allegedly said e.t.c.
I have to agree with those in the industry who are becoming more and more alarmed about the prospects this technology offers. As with all technologies, it is not the technology itself which is evil, it’s the fact that mankind will misuse it, as it has misused anything given to it since the invention of the hammer.
At work some of our tools these days show AI generated “suggestions”. If I only imagine what would happen if we took them on board without strict checks and balances, the result would be disastrous and that is only for weather forecasting…
Imagine this kind of capabilities in the hands of the likes like populist politicians for campaigns which even now are full of lies and distortions. What if you end up getting “documentaries” full of top notch trust bearers, none of which have actually taken part but were generated artificially? Putin could generate “speeches” by foreign leaders who never made them and incense his people, imagine what the Brexit campaign could have been like if you get her late and lamented Majesty to make comments she never would have uttered, but which were fed into her Avatar by Farrage or the likes? Imagine what someone like Bannon or radical environmentalists can make out out this?
With the generic trust into governments and organisations at an all time low since the middle ages, this technology could indeed do exactly what those people who warn from it from the inside fear. Very much so.
If you are following the OpenAI fiasco, it looks like we are in for some very surprising developments. The good news is at least today we can still unplug things. Not sure how long that will be the case. It’s going to be a very interesting few years.
eurogaguest1980 wrote:
The good news is at least today we can still unplug things.
Can we really?
Not so sure. Try to do a few days of simple everyday work without your phone… maybe you could, I would fail miserably. Hence I always carry a back up, like with my glasses as well.
eurogaguest1980 wrote:
It’s going to be a very interesting few years.
Yes, the kind of “excitement” I would love to do without.
The most dangerous streams usually emerge out of things which got underestimated. That is why I am very careful not to fall into that trap.
It is not difficult to identify malicious posters when they join up. Obviously I am not going to post what I do to achieve that. But very very few have got through the “filter”. And actually most if not all of that “filter” could be automated.
The harder bit is when a human joins up with the objective to litter our place with ex chatgpt crap.
It will be more fun if the said human joins up and then uses a bot to do subsequent posts. I see this quite obviously happening on FB (which is an obvious bot posting target, along with twitter etc) but is unlikely here due to the custom software. And there is a bit of time investment to create each new user, needed to enable a bot to do subsequent posts.
It will just trash most other forums out there, assisted by the mods mostly not giving a sh*it because they want activity first and foremost.
The best defence from AI-generated crap is your brain. You cannot fix the fact that a lot of people don’t have one, but those who do, and have a critical mind, will be just fine.
Mooney_Driver wrote:
Yes, the kind of “excitement” I would love to do without.The most dangerous streams usually emerge out of things which got underestimated. That is why I am very careful not to fall into that trap.
I could not agree more. The improvements in AI are going to be exponential.
About unplugging, today I can go into any of our datacenters and power off any machine. Access to these facilities is mostly done by humans. When we turn over physical access control to systems we no longer understand, we’ll have a problem. AIs are already generating code we can’t understand. We don’t even understand how ChatGPT works, really. Every test giving AI “free rein” to learn and produce information has turned out very nasty.
As an aside, I know that AIs today are “just” LLMs that digest and send out information based on understood rules, but this is quickly changing.
Peter wrote:
It will just trash most other forums out there, assisted by the mods mostly not giving a sh*it because they want activity first and foremost.
There is not a lot left to trash really.
First of all it is quite difficult to find mods today who are worth anything and willing to spend the time doing it AND who are knowledgable in the field the forum is all about.
Secondly, most fora have degraded to a few die hard left over old folks who bemoan the fact that fora have been largely replaced by social media, at least where posts about the positive sides of flying and other activities is all about. Travel reports, e.t.c usually end up either as VLOGS on youtube or as serial posts on FB or Twitter or Instagram and similar.
What happens if activity is all they are after can be seen in some prominent places: If you don’t have mods who are after activity over quality, the off topic sections will thrive and the expert sections die, with the experts going to friendlier places. Happened a lot in aviation recently. And that is why most people in Europe come here.
