Recent self-driving car incidents highlight safety concerns

Driverless cars may indeed be the future, but right now, experts are concerned about the dangers they pose to road users. 

In a recent incident in San Francisco, an autonomous car owned by General Motors subsidiary Cruise slowed down a fire truck responding to an emergency. Wired reports that the fire truck was attempting to bypass a double-parked garbage truck by using the opposite lane when it encountered the driverless Cruise vehicle. The Cruise car did not move out of the way as a regular car with a driver would know to do. Eventually, the fire truck got to its destination, but apparently with significant delay. The San Francisco Fire Department said that the incident slowed their response and resulted in property damage and personal injuries.

This incident comes as a state agency is defining rules for Cruise to operate limited but paid ride-hail services across California. A spokesperson for Cruise confirmed the incident but refuted the length of the delay. According to the company, remote assistance workers took control of the car and moved it out of the way, resulting in just a 25-second delay for the fire department. 

In the autonomous driver industry, this kind of occurrence is known as an “edge case,” a rare event that can be difficult to train a car’s AI to respond to, compared to something like stopping at a red light. Even so, this isn’t the first controversy or “edge case” involving driverless cars in recent months. 

Pony.ai, a self-driving startup with funding from Toyota, recently lost its California permit to use safety drivers to test self-driving vehicles because many of the drivers had multiple driving violations. According to Jalopnik, the revocation came after one of the company’s cars collided with a line divider at the end of 2021. An investigation by the DMV followed. Even the most famous champion of self-driving cars, Tesla, has not avoided controversy. The Los Angeles Times reports that a recent documentary called “Elon Musk’s Crash Course” explores the company’s lax safety culture and the safety regulators that allow it. Additionally, a recent study by researchers at the University of California, Irvine, found that autonomous vehicles can be tricked into halting abruptly if someone places an ordinary object — such as a box, bike, or traffic cone — on the side of the road. This is cause for concern as such an abrupt stop could cause a serious accident. 

A report from The Guardian discusses how so-called edge cases aren’t actually all that rare, quoting Melanie Mitchell, a computer scientist and professor of complexity at the Santa Fe Institute. Mitchell says that edge cases “might be infrequent for an individual driver, [but] if you average out over all the drivers in the world, these kinds of edge cases are happening very frequently to somebody.” Mitchell then expresses concerns about the feasibility of training AI to have common-sense responses to edge-case incidents when we don’t yet completely understand how such responses work in humans. 

In other news

• Remote learning software used children’s data for targeted advertising. According to Gizmodo, a report from Human Rights Watch says in 49 countries studied between March and August of 2021, 146 out of 164 Edtech products gathered data or tracked the activities of children when not using the app. Researchers found that in Texas and California, videoconferencing programs not designed for children’s use, including Zoom and Google Meet, were used for remote classwork. Some of these allegedly track the data of their users. Other programs designed for children, like ST Math, used keylogging tracking technology to gather data and send it to third parties, like Google, Facebook, and Spotify. 

• Google Colab bans deepfake training. Over the last two weeks, the Google Research browser platform used for writing and executing python code has changed its terms to disallow the creation of deepfakes. Unite.ai reports that although Colabatory has changed its position on deepfakes, warnings are generated but the code can still be executed. Of the two controversial deepfake creation distributions, only DeepFaceLab has been targeted, and not FaceSwap. A statement from co-lead developer at FaceSwap, Matt Tora, suggests that rather than ethics, Colab’s decision was based on the high amount of resources and time spent on deepfake creation. It is still not clear what other activities may be prohibited in the new changes, such as crypto mining or file-to-file sharing. 

• Universal Music Group to release NFTs on LimeWire. A new NFT marketplace from Limewire on Algorand (a proof-of-stake blockchain network) has just signed Universal Music Group to the platform, according to Decrypt. The major music label represents artists including Kendrick Lamar, the Rolling Stones, and U2, who will all have the chance to release NFTs on the Limewire platform. This will typically be for music, artwork, and related items where the NFT represents proof of ownership. In the same week, Napster also announced its plans to launch an NFT marketplace. Both Limewire and Napster were peer-to-peer file-sharing services in the 2000s.

• Need to power your IoT device? Try algae. A new system has been developed to continuously power a microprocessor for six months with cyanobacteria, according to Smithsonian Magazine. The research from Cambridge University’s biochemistry department was published in Energy and Environmental Science. A small container of the bacteria was placed in a window because it generates energy through photosynthesis, though the process continues in the dark and an electrical current is still generated. In the experiment cyanobacteria powered an ARM Cortex M0+ microprocessor, which is used for various devices in the Internet of Things category. As IoT is a rapidly expanding market, this process could be a more sustainable energy source than lithium batteries. The Verge suggests that it would take 333,000,000 of these algae units to power a desktop computer, but the research team says the system can still be improved.         

• Twitter to pay $150 million penalty for failing privacy. The US Federal Trade Commission (FTC) reports in its business blog that Twitter has been asking its users for personal information for securing their accounts, then uses the data for advertising purposes. In 2010 the FTC cited Twitter over message privacy, and the social media company agreed to an order that would lead to heavy penalties for any future violations. The Department of Justice announced the civil penalty last week after Twitter used multi-factor authentication data in targeted ads from May 2013 to September 2019. A new order has been created to ensure the social media company is more transparent and will not continue to misuse user data.   

• WordPress releases version 6.0 and turns 19. WordPress is celebrating 19 years since Matt Mullenweg and Mike Little released the first version of WordPress on May 27, 2003. The birthday celebrations include tributes from the founders and WordPress enthusiasts on the specially designed website, wp19.day. A few days before the anniversary, the content management system’s latest version, “6. 0 Arturo”, was introduced. Search Engine Journal reports that Arturo brings almost 1000 updates, which include faster website performance, page creation patterns, and global style variations. 
  
  Also, Namecheap’s Managed WordPress Hosting, EasyWP,  turned 5 this week, and customers can get up to 65% on hosting plans until June 6.

Tip of the week

Driverless systems today are gaining popularity, but with rising safety concerns, they have yet to fully supersede a human’s seat behind the wheel. Autonomous cars are now able to shift gears, follow GPS systems, and even use brakes. Despite the big leaps forward with modern-day inventions such as these, car accidents remain a leading cause of injury in many countries around the globe. 

Formula 1 lifestyle website Paddock has a few great tips on staying safe while driving a vehicle with autonomous driving functions: 

• Avoid overreliance on the system and remain focused. 
• Be sure that you are following the speed limit
• Trust your own instincts if your car’s warning system sends an alert that doesn’t seem right. 

It’s also a good idea to check your car’s battery on a regular basis to avoid electrical issues which can send on-board systems into a frenzy.