The major social networking site Facebook is reportedly developing new software that will detect whether a photo contains potentially humiliating elements.
The algorithm is being designed to scan photographs and pick out any features that may be considered embarrassing to the poster, such as whether they appear to be inebriated in the snap. It would then ask the user if they want to continue with the upload.
With many images of drunken nights out being regularly splashed across news feeds, often to the humiliation of users once they have sobered up, Facebook is hoping that it may save people from not just embarrassment, but potential mistakes.
The social firm’s artificial intelligence (AI) research lab is leading the initiative, with its chief, Yann LeCun, telling the magazine Wired that the software would act as an “intelligent digital assistant”. LeCun added that with many Facebook members having relatives and work colleagues, including bosses, on their friend lists, the prompt from the service could help them to rethink their postings.
As people have in the past lost their jobs over something inappropriate being posted on social media, the new algorithm may prove useful to some users.
Reportedly, the ‘deep learning’ software will eventually not only look at photographs, but other posts too. This field of AI has seen a lot of attention and investment as of late, not just from Facebook, but other major online firms as well.
For example, Google recently purchased Deepmind, a start-up focused on deep learning AI. The UK-based company has since entered a partnership with Oxford University.
- Instagram to expand advertising options - June 8, 2015
- A week of Twitter mishaps - June 5, 2015
- Twitter unleashes new analytics tool - June 4, 2015
