UKH

Every now and then, I do have a peep at Rocktalk etc and don’t want to add to the derailment of that post, but:

As others have said, we do get bots and god knows what else on here. Only last week, a bot/imposter posted early in the morning, with a username about one vowel different from the username of a regular poster (I reported it and it got deleted).  I reckon this must happen all over the internet. Every time one of those odd bot types have posted on here, trying to sell Tupperware or asking for the best Trail to Hike, they have zero user profile. It would make sense for new posters to do a quick profile. Takes 30 seconds, and may reduce hassle and embarrassment.

Edit: especially when a new poster posts their first post immediately after creating a new profile. Theres enough negativity kicking around without giving fellow climbers a load of grief, when it is avoidable. 

Maybe there should be one of those prove you're a human typed things when you create a new account, but like a climbing specific one.

Perhaps new users should have to answer some simple questions:

'What have you ever done on grit?'

'What grade is three pebble slab?'

Show some pictures of snow covered crags and pick which ones are 'in'.

Obviously I'm not being entirely serious with those, but I have seen that sort of thing on other forums, answer a simple questions to try to prove you're a genuine user. Maybe UKC has this but it's too long since I created an account to remember. Maybe bots can deal with that kind of thing these days anyway?

I read recently that bots are now better at completing Capcha tests than are humans!

I’ve heard similar, but that bots are more likely to include a typo. BTW, it’s Captcha

It’s a pity that cynics like me, when we see the green L plate sign next to a users name, check the date when the user registered, and if this is matched to an unusual post then our Bot Alert system kicks in.  

It was I who kicked off the Bot speculation on the other thread, i'm sorry for that, I was clearly wrong.

In the reasonably near future I believe it's going to be impossible to tell what, on the internet, is AI generated and human generated.  I'm not sure what point i'm making here, other than I keep turning up my 'is it a bot' detector and if as a UKC community we want to maintain interesting and useful human-human discussion there are going to have to be some changes in how stuff is moderated.  Having a requirement to create a profile, or a cooling off period after registering would seem like a good start.

There's a very simple solution here. If you see a post that you think is a bot, then report it to the moderation team and let the process already in place do it's job. Don't jump in posting accusations at potential new members. Not rocket science. 

Given a large amount of people don't know the correct answer, maybe not the best question  

I'm not sure either of these would help all that much. If the bot can generate plausible posts, it's not going to be much more difficult to generate a plausible profile. The concept of a detailed profile is quite unusual in online forums anyway, so lots of genuine new members might balk at being forced to fill it in and either not finish registering or fill it in with the bare minimum of text to satisfy the requirements without actually giving any useful info.

And a cooling off period isn't very off-putting to an automated system, but would be aggravating to genuine new members with something they want to post. It would give moderators more chance to remove the bot before it posts, but that's only any use if it can be identified as a bot without the benefit of any posts to show its intentions.

Both steps might well put off more genuine people than bots. We're at the point where there's genuinely significant overlap between the "writing" capabilities of people and LLMs, and that can only get worse from here. It might be that moderation soon has to let go of caring whether a poster is a person or a bot and just focus on whether or not they're making a constructive contribution. If someone joins, posts one vaguely plausible climbing question, then starts spamming disinformation and conspiracy theories, it might be hard to tell whether they're real or a bot, but either way I'd rather see the back of them so identification ultimately doesn't matter. Maybe you'd get some bots flying under the radar just chatting nonsense about the grade of TPS, but the incentive to run bots is normally to push an agenda, so that should give them away most of the time, even if it only identifies them as a malign influence rather than ever knowing whether it was a bot or a person.

+1 to this.

A while back we had a person that would regularly create a new profile, do one post about climbing and then launch into conspiracy theories. It was almost certainly one or two people, real people, who were climbers. There is no way to automatically exclude them without excluding bona-fide users. All we can do is allow free access and then remove any user that abuses it. Which is what happens now, and very efficiently. I regularly report spam / bot threads and they disappear quickly.

What we really don't want to be doing is laying into a new user accusing them of being a bot like on that other thread. I have to admit I thought the OP looked suspicious, but I didn't think it was reportable so I just ignored it. The way that thread went is just embarrassing. It's the last thing the OP deserved, but exactly what a bot would have wanted!

https://xkcd.com/810/

On this theme this made me laugh a lot when it came out a while ago

https://youtu.be/LButXcZ57pc?si=a3w-BFSYtuvidS8-

I agree with others about the rise of AI/Bots in terms of content generation, feels like ‘the internet’ might be at a tipping point where it ceases to be a useful tool expect for very heavily moderated/controlled sites…..Google searches are already proving a lot less useful personally for me that they used to

I'm not sure how that would improve things.

Your own profile has a lot of the characteristics of something created by a bot.

Other forums require new users to post a certain number of times before they can post links. Then they have to type out their conspiracy theories instead of link to them. 

Have to admit, out of all the varied forums I follow UKC is the only one I've seen obvious bots on....and the only one where anyone new is accused off being one.

There does seem to be a particular issue with UKC, much more than any other forum I frequent. 

I wonder if it's been picked on as a good bot training resource?  If your bot can make it here, it's probably good to be used anywhere!

It’s even odder given ukc have a custom forum software and not an off the shelf one.

Name checks out...

There's an arms race between spammers and defenders. It used to be that bots were dumb, so setting a simple JavaScript challenge like 2+2, would stop them. Now, some bots will be able to answer the many background challenges, before you get to the 'how many motorbikes' pics. The cat and mouse has matured to include humans in the chain. There is an example in the link. It seems like a nice work from home job.

https://www.f5.com/company/blog/how-cybercriminals-bypass-captcha

Why are ukc being targeted?

Different resources have different value to the spammers. FB, Twitter, insta accounts are cheap as chips. Resources with different audiences are ratted according to potential impact. So a global audience, with a lot of different professions, different levels of professional etc might be more useful in spreading certain styles of content than others.

Who is doing it?

Every person and their dog, the barriers to entry are pretty low. Viral marketeers, hyping product, would be the bottom branch. Going up to paid for influence / disinformation campaigns. With state sponsored campaigns at the top.

Why?

We saw with Trump, it was far cheaper and more effective to elicit regime change, via facebook than other methods. Normally overthrowing a government and installing a favourable leader, might involve tanks on the street and such. I expect that certain celebs, will be looking at damage limitation etc. In order for these campaigns to be successful, you need the infrastructure in place. Embedded accounts, across a multitude of platforms, with trust in place and ready to go. You cant take a month to spin up 30K accounts, if you need to react to a breaking story.

Damn, I’ve been sussed.

what do you keep your sandwiches in?

Funny you should mention that. For years I’ve struggled with this, using tin foil, Christmas wrapping paper and old y fronts. Squashed butties, mangled pies - made my life hell. But a friend showed me ToupeeWear. It’s a hair piece that you wrap your sarnies in. Also great for Wigan Kebabs, Chips and Pea Wet and all manner of healthy tasty lunchtime goodies. www.ToupeeWear.com.co.org