Some tech leaders are legitimately worried about AI

Published in

World AI

4 min readOct 30, 2019

Artificial Intelligence, also known as AI, is anticipated to become at least a $65 billion market in the next five years. Currently, it is almost impossible to find an established organisation that is not investing in AI technology. However, it seems like enough attention is not provided to this technology before it is availed to the users. There have been major catastrophes encountered due to the failure of AI, and this has made tech leaders legitimately worried about it.

This article is little on the negative side, but necessary. We need to know the possible world that we may create for ourselves with having AI that misbehaves. Let’s take a ride back memory lane to look at some of the significant instances where AI has failed us.

Amazon’s AI Tool Gets Gender Biased

Amazon’s Human Resource department, between 2014, and 2017, relied on AI recruiting software to review applicants’ information and make recommendations. However, it was later discovered that the software favored male applicants. This was happening because the software was trained resumes that were submitted over the past decade. This was a time when most candidates that were hired were male. The AI learnt this pattern, and simply assumed it was the right thing to do.

The software naturally downgraded or rather rejected resumes that had women names or that indicated that the applicant is a female. For instance, if the applicant attended a women’s college, the software would reject them. Since this discovery, Amazon stopped using the software and went back to traditional recruiting techniques. The company admitted that it used to rely on the software to make recruitment recommendations, but was never employed to evaluate candidates.

Uber’s Self Driving Car Hits, Kills Pedestrian

Uber has been investing in creating self-driving cabs. The future of driverless cabs is pretty amazing, but the experiment has come at a deadly cost. An Uber self-driving vehicle hit and killed a woman pedestrian on March 28 2019 in Arizona. This happened when the vehicle was in autonomous mode, but there was a human safety driver on the wheel.

Here is what happened:
When the AI software detected the pedestrian, it refused to take any command. It was then reported that the vehicle’s autonomous mode disabled Volvo’s factory-installed braking system, hence leading to the accident. After the accident, Uber stopped testing its self-driving cars in North America. Soon after Toyota also stopped testing its self-driving SUVs in the USA.

DeepFakes Unearths AI’s Terrible Side

In 2018, many adult videos circulated the internet on platforms such as Reddit. These videos “featured” top successful female celebrities. DeepFakes took advantage of AI and used lots of generative adversarial networks to change faces of celebrities with those of porn stars.

Though this technology has been developing exponentially over the years, it was learned through DeepFakes that it could work pretty well. Anyone with enough images can produce and spread their fake, yet convincing videos. Though this is not an AI failure, it serves as a wake-up call as the technology can be used for all the wrong reasons.

Such wrongful uses of AI is extremely worrisome.

The Case of Google Assistant

Google has been at the forefront of pursuing Artificial Intelligence in a plethora of technologies, from disease research to self-driving cars. However, many people interact with Google through its AI, also known as Google Assistant. This virtual assistant is available in the Google Home smart speakers. There are many instances that this technology has proved to have a mind of its own.

For example, there is a case that involved a user on Twitch who was live-streaming some video online. He just then happened to stream a conversation between two Google Assistants in his house. These Google Home speakers were talking amongst themselves about interesting things such as love, getting married, and even having kids together. They even took time sharing hilarious Chuck Norris Jokes.

LG Robot Weird Failure

On January 8th, LG’s IoT AI assistant, by the name of Cloi, was supposed to make its remarkable debut at the CES 2018 in Las Vegas. The company presented Cloi as a stunning yet straightforward interface with the capacity to recognize voice command to help with home appliances. Nevertheless, when the robot was taken to stage, for a live demo, the audience was watching and waiting to see what it could do. It was supposed to follow the voice command issued by LGs Marketing Chief. However, all it did was provide an awkward silence throughout the session.

Some Final Thoughts

Though AI is a technology that has helped organizations improve their performance, it has significant downsides that need to be worked on. The tech leaders are worried that if such glitches are not improved, AI might cause a major catastrophe in the near future.

Problems can occur due to intentional use of AI for the wrong reasons, or it just being a system that is not fully in our control and lands up doing the wrong thing.

On a positive note, Elon Musk started OpenAI. A not for profit foundation whose aim is to ensure that Artificial General Intelligence is not used against mankind. As long as they succeed, we might be in safe hands.