13/01/2015 06:29 GMT | Updated 13/01/2015 06:59 GMT

Computers Know You Better Than Your Mates Do

A computer can be a better judge of character than an individual's parents or close friends, research has shown.

All it needs is the right input data - namely the thumbs-up clues left by someone's Facebook Likes. By "mining" for Likes, the software was able to predict personality more accurately than friends and family.

Only husbands and wives rivalled the computer's ability to sum up broad psychological traits. (Something that may worry our tech reporter Tom Tamblyn based on this blog...)

The research is an "important milestone" on the path towards more natural and social interactions between humans and computers or robots, according to the scientists.

Lead author Ms Wu Youyou, from Cambridge University's Psychometrics Centre, said:

"In the future, computers could be able to infer our psychological traits and react accordingly, leading to the emergence of emotionally-intelligent and socially skilled machines.

"In this context, the human-computer interactions depicted in science fiction films such as Her seem to be within our reach."

In Her, starring Joaquin Phoenix, a man develops a close relationship with an intelligent computer operating system personified by a female voice and calling itself Samantha. But the scientists admit there may be concerns about privacy as such technology develops, and say they support policies giving users full control over their digital footprint.

The researchers found that their software was able to predict a study participant's personality more accurately than a work colleague by analysing just 10 Likes.

Inputting 70 Likes allowed it to obtain a truer picture of someone's character than a friend or room-mate, while 150 Likes out-performed a parent or sibling.

It took 300 Likes before the programme was able to judge character better than a spouse. Given that an average Facebook user has about 227 Likes, the researchers say this kind of artificial intelligence (AI) has the potential to know us better than our closest companions.

SEE ALSO:Researchers Say AI Inventors 'Must Avoid Apocalyptic Mistakes'

The work, reported in the journal Proceedings of the National Academy of Sciences, builds on a previous study by Cambridge University scientists which showed that a variety of psychological and demographic characteristics can be predicted through Facebook Likes.

For the new research, 86,220 volunteers on Facebook completed a 100-item personality questionnaire and allowed their Likes to be accessed.

The results provided self-reported personality scores for what are known as the "big five" psychological traits: openness, conscientiousness, extraversion, agreeableness, and neuroticism (Ocean).

Analysis of the data showed which Likes equated with higher levels of particular traits. For instance, liking "Salvador Dali" or "meditation" revealed a high degree of "openness".

The results of the computerised character assessment were compared with judgements of friends and family members made using a shorter version of the personality test.

Given enough Likes, the software matched people's self-reported personality traits more closely than siblings, parents or partners.

Potentially the technology could influence who we employ, elect, or even marry, say the scientists.

Co-author Dr David Stillwell, also from Cambridge University, said: "The ability to judge personality is an essential component of social living - from day-to-day decisions to long-term plans such as whom to marry, trust, hire, or elect as president.

"The results of such data analysis can be very useful in aiding people when making decisions."

Phd student Ms Youyou added: "Recruiters could better match candidates with jobs based on their personality; products and services could adjust their behaviour to best match their users' characters and changing moods.

"People may choose to augment their own intuitions and judgments with this kind of data analysis when making important life decisions such as choosing activities, career paths, or even romantic partners. Such data-driven decisions may well improve people's lives."

But the researchers share the concerns of those who fear a dystopian future in which our traits and habits become an "open book" for computers to read.

Dr Michal Kosinski, another member of the team from Stanford University in the US, said: "We hope that consumers, technology developers, and policy-makers will tackle those challenges by supporting privacy-protecting laws and technologies, and giving the users full control over their digital footprints."