Liveblog of UK BlueLightCamp Unconference session on public sector monitoring of social media. Probe to errors, uncertainly and howling cries against grammar.
The UK Snow Map – reveals (to a degree) which Twitter users live where, based on the first part of their postcode.
Another example – when people were tweeting about a big night out in Guildford, the police responded saying – we’ll be there, have a great time.
They use TweetReach, constantly looking for mentions of their major place name. They have a crap data problem, with people tapping “woking” rather than “working”, or “epsom salts” tweets coming up instead of Epsom tweets. That has data over year old, now – is that a bit “big brother”-ish?
If they’re monitoring mentions of Guildford, and somebody tweets that they’re going to commit suicide there, and they miss it, is that icky?
Why are you collecting the data? If you have a clear reason, it becomes easier. Accidental creation of datasets without a clear policing agenda certainly brings a cringe factor.
Scanning Twitter and intervening during the riots seems proportionate. Is it the same doing that on an average Saturday night? We expect the police to monitor and gather data – we don’t expect that so much from local authorities.
Have they done social network analysis around the postings identified? No.
Isn’t there a duty to influence some groups? Monitoring and intervening is a huge ethical issue. Is it better to identify the community, and find an influencer to spread the message. Is it creepy is you send a message directly, or is it more creepy if you place a message where they’re looking.
There’s a difference between butting in to give information they want, and interrupting to warn or send a negative massage. As a society we haven’t yet developed a language for understanding these public social networks. Why do we assume that it’s OK to treat messages published on Facebook as public?
The state has more powers to constrain liberty than a company, so we need to be more careful with this. We’ve been seduced with the fact we can do it. Do we end up advertising the fact we’re monitoring by examples like the Guildford Saturday night tweets?
Waze – the crowd-sourced navigation app – tells you when traffic is building up ahead, and where the speed-cameras are. And the latter is technically illegal.
The social context of monitoring
Is the police monitoring and responding to what you’re doing a human rights abuse, if done in real time? It’s very different from analysing data afterwards. To take that further, if people tweet about meeting up and doing things – and then the police get there first. Can you intervene? Yes, and charge them with potential breach of the peace.
Do you look at context with a tweet. For example, a tweet about bombing the airport – do you check the conversation to see if its part of a messing around conversation?
At the moment, you’re asking “should we be doing this?”. The other angle is asking if people are expecting a mention of an organisation to be followed up? Over 70% of younger people expect customer service bases on a Twitter arrangement. What is this “virtual contract” we have?
The icky bit is the assumption that you’ve put everything on a public social site, therefore you have no privacy. Are Facebook Pages a better place to intervene, because they’re a public space, than someone’s wall, which feels more “private”.
But… social media tools provide you with the ability to lock down your profile. That doesn’t mean everyone does.
Terms of engagement
Has any organisation published what it plans to do with Open Source Intelligence? Could they do that? Set rules of engagement? Can those change in terms of emergency?
Could you publish your search terms? No great, obviously, for policing, but maybe for other bodies. If you do say, you have an issue if you miss something.
Should we take a forum moderation approach? You just collect the data, and only interrogate it when something is brought to you attention?
Crackdown – a single place to report anti-social driving. Do we need something like this for stuff spotted on social media?
Firefox was doing a poll about what you want from the web. Pretty much everyone in the results asked for privacy. The general move is back towards private service like Snapchat and the like. Seeing people’s Snapchat feels more icky than their tweet. There’s no doubt that the short term trend is towards privacy, but the long term trend still seems to be towards openness.
Post-Snowdon people are less trusting of what we way we’re doing with information. If you have the data, the justification is to serve and protect the public. Easier for the police officer, than for the marketing and comms team…