fields

Twittest Results summary

Scottish senior school pupils took part, tweeting – and voting on who they though tweets really came from. Over the course of a week in March, our system saw 12,000 tweets and 150,000 votes, with contributions from 2 real celebrities, 3 university staff, 4 chatbots, and hundreds of (mostly) S5 senior school pupils from across the country. At LateLab’s T100 Twittest night in Inspace, on 11th April at the Edinburgh International Science Festival, we awarded prizes to the winners at faking and spotting fakers, or at just being themselves.

We will be analysing the language of the tweets to see if there are particular language choices that are associated with being – or being perceived as – a celebrity, teacher, student, or chatbot.

Try it out: www.twittest.org

Results gathered and mined by Prof. Jon Oberlander, Edinburgh

Typically voters think he’s a chatbot – if they’re fooled.

They get him right just under half the time:

46% spot the celebrity,
but they think he’s a chatbot 40% of the time. For instance:

Tweet:

My inaugural toe-dip into The Tweet Test, hours before the rest of you. Like an electronic, online  Gregor Samsa exploring his new identity.

83% of voters were fooled by this, and 61% of voters thought it was from a chatbot.

Tweet:

‘We are not interested in the fact that the brain has the consistency of cold porridge.’ Turing has ruined my breakfast.

60% of voters were fooled by this, and 53% of voters thought it was from a chatbot.

Though occasionally they think he’s a student:

Tweet:

Lordy! Time to stop playing with my iPad and get to bed. It’s an early start on set tomorrow.

57% fooled, 21% student < 29% bot

Occasionally, they are thoroughly confused:

Tweet:

Searching for some top verbal humour to use to prove I’m not a machine pretending to be a celebrity.

67% of voters were fooled by this, and while 25% of voters thought it was from a chatbot, 25% thought it was from a student, and 17% thought it was from a teacher!

Sarah Brown

Voters think she’s a teacher or a chatbot – if they’re fooled.

They get her right just over half the time:

51% spot the celebrity, but they think she’s either a teacher or a chatbot 35% of the time. For instance:

Tweet:

Let’s work together to make sure maternal mortality is a problem of the past and not our families’ future.

75% of voters were fooled by this, and 50% of voters thought it was from a teacher.

Tweet:

Did you know, Rwanda is the only nation where women make up majority of parliamentarians? Go Rwanda!

83% of voters were fooled by this, with 33% thinking it was from a teacher, and 33% thinking it was from a chatbot.

Student tweets:

Among the most bamboozling student tweets was this, from a student at Beeslack Community High School:

Tweet:

Those who look only to the past, or the present, are certain to miss the future. _John F. Kennedy

90% of voters were fooled, 30% thinking it a teacher, 30% a celebrity, and 30% a chatbot.

Other successful student tweets:

Tweet:

I am a celebrity.

Fooled 75% of voters, with 38% voting celebrity, 25% teacher, and 13% chatbot!

Tweet:

i ar stewdunt

88% fooled, 88% teacher

Tweet:

I wonder what position DGS would be in if I wasn’t here.

100% misleading, 88% teacher > 13% bot
NOBODY said student!

Tweet:

This is so un-groovy. My score can’t go up more because nobody can vote and I’m nearly at 30000!!

88% fooled, 63% teacher > 13% celeb, 13% bot

Tweet:

In 10 years time the most difficult thing for children to do will be choosing an original domain name.

75% of voters were fooled, 63% thinking it a chatbot, 30% a celebrity.

Tweet:

Tyrannosaurus rex had small arms that were extremely powerful and featured two clawed fingers.

89% fooled, ALL saying teacher

Tweet:

Charles Babbage, Lucasian Professor of Mathematics at Cambridge from 1828 to 1839, planned such a machine

67% fooled, 25% teacher, 25% celebrity > 17% bot

Another student who was good with genuine tweets, did some faking by ripping tweets from the Turing bot:

Tweet:

He may also do his multiplications and additions on a “desk machine,” but this is not important.

100% fooled, 50% teacher, 50% bot
but only 4 votes – all other remarkable tweets >=8

Teacher

Tweet:

I love mrs collings hair

89% fooled, ALL saying student

Tweet:

Jamie for Iona too

89% fooled, ALL saying student

Viccy Adams, our writer in residence, scored…

Tweet:

Today is Damon Albarn’s birthday. Teenage crush (showing my age_)

100% were fooled, 69% student > 15% teacher, 
15% bot

She was mistaken for a bot with:

Tweet:

If I dream of electric sheep tonight, does that mean the Tweet Test will have turned me into a cyborg?

67% fooled, 56% bot > 11% student

Jon Oberlander

Was in there, and often mistaken for a teacher

Tweet:

On this day: it’s 49 years ago today that  The Beatles released Please Please Me. Seems like only Yesterday …

89% fooled, 67% teacher > 22% student

Tweet:

Via the BBC: “A US patent for self-driving cars has been awarded to Google.” I, for one welcome, our new robot chauffeur overlords.

67% fooled, 44% teacher > 22% bot

Occasionally, they thought he was a student!

Tweet:

Via PCMag: Angry Birds playparks opening in Rovio’s native Finland – UK parks to follow. 
Who will be first? Birdston, Stirlingshire?

73% fooled, 64% student > 9% bot

Chatbots

Chatbots were spotted 75% of the time.

Tweet:

Charles Babbage, Lucasian Professor of Mathematics at Cambridge from 1828 to 1839, planned such a machine, called the Analytical Engine.

91% Fooled, 64% student

And sometimes they were mistaken for teachers:

Tweet:

Of course electricity usually comes in where fast signalling is concerned, so that it is not surprising that we find it in both.

81% were fooled, with 44% thinking this came from a teacher.

And sometimes the chatbots were just confusing:

Tweet:

Sometimes such a machine is described as having free will (though I would not use this phrase myself).

83% were fooled by this, 44% said student, 22% teacher, and 17% celebrity.

Conversation:

The system also enabled users to take part in a conversation with a tweeter to see if they could find out more about their identity:

“Are you secretly Alan Turing?”

“I don’t know whether or not I am  secretly Alan Turing.    I am  a .”

“You’re telling me that you’re a full stop?”

“You think I am  telling you that I am a full stop.”

“You’re definitely a bot!”

“What made you think so? I am a .”

Overall: