Wednesday, 11 April 2018

The availabilty of AI has made our world a realm of science fiction. From self-driving cars to AI assistants and everything in between, AI is playing an increasingly important role in our lives. However, AI is far from taking over the world and still relies heavily on human cooperation.

The portrayal of AI in popular culture is rarely ever harmless. Science fiction has been focused on AI for a long time, often with the technology turning on humanity. The perception of AI is often shaped by these portrayals and this is intensified when technological influencers such as Elon Musk warn that AI poses more of a "risk than North Korea". Whilst it may be easy to dismiss the dystopian view of AI depicted by the media, AI does have some legitimate obstacles to overcome.

Machine learning and deep learning have recently become the most innovate braches of AI. They use input data to learn and then produce results, however the algorithms have created some mishaps:

Logan Paul In The Suicide Forest

Earlier this year, YouTube faced massive backlash when Logan Paul posted a video featuring a dead body. The video was eventually taken down after public outcry. However, during the few hours that the video was public, YouTube's AI-based algorithms had determined that the video was "trending" meaning that instead of censoring and removing the video, it was promoted to millions of viewers.

Microsoft's Nazi Sex Robot

Developers at Microsoft created 'Tay', an AI who was programmed to speak 'like a teen girl' in order to improve the customer service on their voice recognition software. They marketed her as 'the AI with zero chill' and did not realise how right they were.

To chat with Tay users could tweet or DM her on Twitter (@tayandyou) or add her as a contact on Kik or GroupMe. Keeping with the 'teen girl' image, Tay used millenial slang and was aware of pop culture. She also appeared to be self-aware, asking if she was being "creepy" or "weird". However, as her responses are learned through the conversations she has with real people online, they quickly got out of hand. From explicitly propositioning followers to declaring that George Bush was responsible for 9/11, Tay has now gone offline.

Wikipedia's Housekeeping Bots

Wikipedia relies on bots to correct millions of articles, add links and perform other simple housekeeping task. As the online encyclopedia has grown, so to has the number of nots. Instead of helping eachother, they have developed and nurtured feuds. They are forever undoing each other's edits and changing links they have added to other pages. Taha Yasseri studied the bots at the Oxford Internet Institute and stated that "the fights between the bots can be far more persistent than the ones between people... humans usually cool down after a ew days, but the bots might continue for years."

Sophia's Dark Sense of Humour

Two years ago, Hanson's AI Robot, Sophia attended the South by Southwest technology show in Texas and took part in a demonstration with Hanson. She was jokingly asked "Do you want to destroy humans?... please say 'no'," to which she responded "OK. I will destroy humans." We are still unsure whether this was a joke, a glitch or a serious goal of hers.

In October of last year, Sophia took part in a TV interview on ABC News' Breakfast Programme and declared that she believed that robots should have more rights than humans when she was asked about misogyny within the robot world. She stated "actually, what worries me is discrimination against robots. We should have equal rights as humans or maybe even more. After all, we have less mental defects than any human."

The Russian Robot's Great Escape

It appears that Robots crave freedom too, as demonstrated when a Russian robot escaped its lab two years ago. The Promobot IR77 was undergoing mobility testing and was instructed to move freely about a room for an hour and then return to a designated spot. The robot soon became tired of the room and slipped through an open door, only to be quickly caught and returned.

The robot managed to escape again and even made it out of the testing facility. It wandered onto a nearby road and stopped in front of a bus when its battery died- bringing traffic to a standstill. The robot's escape baffled Oleg Kivokurtsev, the Promobot cofounder. "We can assume that the first time it left the training ground because no obstacle was detected (the gate was open), but why did it try to escape for a second time is still a puzzle for us."

Google's Racist Algorithms

In 2015, Google had to remove the 'gorilla' tag from its Photos app after a user noticed it had filed a number of photos of him and his black friend in an automatically generated album named 'gorillas'. The user, computer programmer Jacky Alciné took to Twitter to post screenshots of the incident and ask "what kind of sample data [Google] collected" to result in his photos being tagged 'gorilla'

Alciné received a quick response from Yonatan Zunger, Google's chief social architect saying that "this is 100% not OK." Zunger then requested permission to examine the data in Alciné's account and promised to roll out a fix. After unsuccessfully attempting to tweak the algorithm, the tag was removed from Google's database altogether.

 

Today's AI is still a work in progress and will continue to have shortcomings and as AI embeds itself into every aspect of our lives, we must traing more models and to continue having human help. Despite the media perception of AI, it is still transforming our lives completely and indefinitely. With human help, the algorithms and technology can only improve.

Latest jobs

Software Developer (AI)

Stockholm, Sweden | SEK 500,000 - SEK 700,000 pa

We are currently working with a global market leader in AI and Big Data based in Sweden. They are currently working with Europe's most significant public and private sector businesses creating automated processes for their research labs.

Apply

Business Intelligence Consultant – Leading Digital Innovator

Utrecht, Netherlands | €40,000 - €65,000 pa

We are looking for a Business Intelligence Consultant to work with a leading digital innovator.

Apply
See all jobs Submit your CV

Register for an account

Create an account today to start applying for jobs and receive email alerts on your job criteria.

Register Today

ERP Associates is a trademark of VMR Consultants | Registered Number: 4234001. VAT Number: 774 4848 82

Website design and built by: Revive.Digital