By Jasmine Gardner/London Evening Standard

Morgan Freeman has died three times: once on Twitter, twice on Facebook. During Hurricane Sandy, a shark swam up the streets of New Jersey. In the London riots of 2011 a tiger was let loose from London Zoo. And last week Manchester United boss David Moyes was sacked.

Of course all this is nothing more than social media rumour - much to the relief, no doubt, of Moyes, who had been tabled by tweeters to be losing his job at 2:30pm last Wednesday, after Manchester United’s 2-0 defeat in the Champions’ League to Olympiakos the day before.

The gossip spread quickly among fans: by midweek, bookmakers had 8/11 odds on Moyes being the next Premier League manager to leave his job. He hasn’t: such are the limits of social media’s power, as yet, to make rumours come true.

Maybe Moyes will be comforted to discover that a new project involving King’s College London called Pheme is attempting, over the next 18 months, to build a web app that would work out how likely a Twitter statement is to be true or false. Until then (perhaps even then, say sceptics) working out whose news you wish to repeat (or gossip you wish to retweet) online will be an ever-growing minefield.

In the US a third of all adults under 30 are said to get their news from social media, with half of all Twitter users receiving their newsflashes in 140 characters. According to FastCompany, social media has overtaken porn consumption in the West as the number one activity on the web - one cheap thrill superseded by another.

More significantly, a survey late last year by market researcher Populus showed that in Britain, Twitter is now an everyday news source for 55% of opinion formers. In other words, a majority of influential people get their information from a potentially unreliable source.

Indeed, last week even talkSPORT presenter Andy Goldstein helped carry the whispers of Moyes’s departure, tweeting “Don’t know how true but NYSE apparently been told Moyes has been sacked”. Who to trust?

Some web companies such as PeerIndex and Klout have attempted to answer this, giving social media users a rating based on their influence across the web. Yet while being trustworthy could enhance your influence, so too can fame, which reduces the usefulness of these services in assessing reliability.

 Ratings go up depending on numbers of followers and re-tweets: according to that maths, we would trust Harry Styles and Justin Bieber above all else.

Yet if trust is in trouble, is that the fault of unreliable people or an unreliable method of communication? Put another way: should I trust the words of someone I follow any more or less were I to meet them in the flesh?

PeerIndex CEO Azeem Azhar once told me about what he saw as the risk of the “filter bubble” of the Internet. His example was related to consumption: “You lose any sense of serendipity because the machines have decided that you’re interested in fishing and romantic comedies so you never get shown the Bourne Legacy or a skiing holiday,” he said.

That applies to online opinion too.

Twitter suggests people to follow based on those with whom we already interact. The more we curate from where and whom we receive information (through selective following and tailored news apps) the less likely we ever are to find out about something outside that remit.

Meanwhile, the web continues to do this for us, algorithmically, with Google showing us personalised search results, based on our browser history, and Facebook editing out the feeds of friends we talk to least.

In his TED talk on the subject, Internet activist Eli Pariser said: “Instead of a balanced information diet you can end up surrounded by information junk food” - where everything we see online is only what we enjoy believing. On Facebook, unsurprisingly, most people’s news consumption comes through friends and family.

In terms of rumour, what we have done (to borrow from the film Meet the Parents) is narrow down our “circle of trust”. It would likely matter little whether this circle was online or off: we begin to think we can rely on all the information we hear from that group.

But this does not quite follow. As Charles H Green, founder of the TrustedAdvisor consultancy, writes, “trust is only weakly transitive. If A trusts B and B trusts C, it follows only weakly that A trusts C.” In other words, I might trust my colleague enough to share her gossip about David Moyes but if she got the information secondhand from someone in her “circle of trust”, then I can only half believe it.

The trouble when you move this model to Twitter is its speed and mass access. Sam Gardiner, the schoolboy who fooled Twitter into believing he was Samuel Rhodes, a football journalist with insider knowledge, observed: “With technology, when people chose to be dishonest they can do it at scale.” Indeed we can - and we want to.

Not only is Twitter a place for influencers to find information; it has become the place for wannabe influencers. Among the general public, just 15% per information from Twitter: so the majority of us bothering to share there have something to promote or a desire to be heard.

And such is the repugnance among tweeps for sharing old information that when we find a rumour, the desire to be the first to retweet it can outweigh the desire to be right. In the meatspace (offline world) my colleague might have passed on her pub tittle-tattle -  but I might have reached home before deciding whether I trusted it enough to pass it on to my partner. On Twitter, we take just seconds to decide whether to spread the word.

Twitter hasn’t taught us to be more trusting or more gullible but simply to speed up our decision-making when it comes to spreading gossip. The danger is we value speed over truth - as David Moyes can ruefully reflect now.

Twitter: @JasGardner