codete the state of self driving cars main 01d87cce18
Codete Blog

The State of Self-Driving Cars in 2023

Avatar Female 7c9e57b38c
avatar male f667854eaa

04/04/2023 |

10 min read

Dominika Reszke,

Piotr Wawryka

In recent years, automated driving has made the headlines many times – sparking admiration on the one hand and controversy on the other. Whenever we heard about AI cars, connected vehicles, driverless taxis, or automated trucks, it was easy to think that the future was now, indeed.

But we’ve also heard about some accidents that raised serious security concerns and questions about the future of self-driving cars. They are being developed extensively, but what might come next? What scenarios can we expect to fulfill this year and beyond?

In this article, we’ll present the state of self-driving cars as of 2023, paying special attention to the role of artificial intelligence in the development of driving automation. On top of that, the 5 levels of autonomous driving will be explained in detail.


Table of contents:

  1. Self-driving cars in brief
  2. The 5 levels of autonomous driving
  3. Autonomous vehicles – benefits and risks
  4. The future of self-driving cars
  5. Driverless cars – key takeaways

Self-driving cars in brief

Vehicle automation, along with vehicle electrification, is an idea that has been rocking the automotive industry for years. Automated driving systems are ones that operate without human intervention or human input, using advanced technologies and AI-based algorithms.

In short, a self-driving car is „a vehicle that uses a combination of sensors, cameras, radar and artificial intelligence (AI) to travel between destinations without a human operator.” Interestingly, the first semi-automated driverless car was launched in Japan as early as in... 1977.

As of now, there are dozens of self-driving car manufacturers or designers in the market worldwide. These include Tesla, NVIDIA, Zoox,, Cruise, Waymo,, Embark, Nuro, Momento, Aurora, and May Mobility, to name but a few notable brands.

Artificial intelligence in autonomous vehicles

One of the reasons why autonomous cars are what they are and expand that fast is Artificial Intelligence. AI lets engineers develop autonomous technologies that can function independently but are fast and flexible enough to respond promptly to various circumstances or events.

Undoubtedly, AI in the automobile industry is truly a game-changer, helping to make things such as autonomous car sensors, actuators, radars, or processors work. Powered by AI, they can be used for data analysis, steering, motion detection, or voice recognition – to make self-driving smoother and safer.

AI algorithms and other technologies, such as image-recognition systems, machine learning, the Internet of Things, and Big Data are used to digest a lot of data, as well as learn and identify recurring patterns and objects, like traffic lights, street signs, and pedestrians.

The 5 levels of autonomous driving

There are several autonomous driving levels, with level 0 standing for no sustained vehicle control, and level 5 – for full automation. This interesting six-level classification system has been developed by The Society of Automotive Engineers and the U.S. National Highway Traffic Safety Administration.

Let’s have a closer look at these levels.

Level 0: Momentary Driver Assistance / No Automation

The driver is in charge of performing all dynamic driving tasks and remains responsible for all aspects of driving a car. Fully engaged, this person is to operate the vehicle on their own with their eyes, hands, and legs. Driver assistance is impermanent and covers brief interventions such as alerts, warnings (e.g., forward collision warning), and emergency safety interventions, like automatic emergency braking.

Level 1: Driver Assistance

In this case (also called the „hands-on” mode), the automated system’s assistance gets a continuous form, although the driver, still, needs to stay fully focused and engaged. The person that’s in the driver's seat can get an AI’s helping hand in terms of either steering the car or accelerating/braking. Examples of level 1 include autonomous parking, adaptive cruise control (ACC), and lane-keeping assistance (LKA).

Level 2: Additional Assistance

This level can also be referred to as partial driving automation, as the system can perform both steering and accelerating/braking. However, it’s still the driver that’s fully in charge of driving, responsible for monitoring the situation on the road and roadside, and ready to intervene promptly at all times if anything goes wrong. For this reason, the driver’s hands should always lie on the steering wheel, despite this level’s „hands off” alternative name. Tesla’s Autopilot feature fits in the level 2 category.

Level 3: Conditional Automation

Here, driving automation gets even further, but it’s conditional; this mode can be called „eyes off.” It’s the system that performs all driving tasks, but the driver needs to be on the alert, too, ready to respond appropriately when requested to intervene, including to take over driving whenever necessary. A traffic jam chauffeur is an example of AV’s level 3.

Level 4: High Automation

In this case, driving automation is high, but not full, yet. A human driver is not necessary for a vehicle to operate, and all people that are in the car can act as passengers only, with a „mind off” – the driver’s seat may even be left empty. It’s the automated driving system that performs all the dynamic driving tasks (albeit in limited spatial areas, as robotic taxis do), and human attention or readiness to take over is not required.

Level 5: Full Automation

This level brings the idea of full self-driving to life, with a „steering wheel optional.” The most advanced self-driving car that represents it doesn’t need human intervention of any kind, similarly as in the case of level 4. The only difference is that here, the automatic system can operate the vehicle on all kinds of roads, in all conditions, all over the world – and not only in certain geofenced areas.

Autonomous vehicles – benefits and risks

Self-driving vehicles make futuristic dreams come true here and now. Being delighted and overwhelmed by these developments, we shouldn’t be uncritical as, apart from many benefits, there are challenges and possible threats involved in the AV trend, too. 

Let’s shed some light on the pros and cons of automated vehicles.

Advantages of self-driving vehicles

Self-driving vehicles seem to be a perfect match for many cases and solve many problems. They may enhance driving in a variety of ways, making it smoother and more optimized – in terms of traffic congestion, the use of infrastructure, and the economy.

Self-driving cars can potentially let societies and governments save a lot of money. That’s because the number of accidents – and their painful and costly consequences – may get trimmed significantly with AV. Some estimations say that the widespread use of autonomous vehicles can reduce the number of motor vehicle crashes by 90%, resulting in thousands of lives and billions of dollars saved.

On top of that, autonomous cars may be the right option for people with disabilities and impairments, the elderly, and those who have difficulties learning to drive. With self-driving vehicles, they would get a great means of transport that’s trustworthy and safe – for themselves and their environment.

Disadvantages of self-driving vehicles

All these advantages of autonomous cars rely upon the undisturbed and failure-free operations of technologically advanced sensors, radars, or cameras. And as we’re talking about traffic and human lives involved, any glitch, error, or bug can have grave consequences.

In an optimal or best-case scenario, real-time communication – embedded in autonomous driving – lets particular vehicles keep the right distance from each other, choose the optimal route, and stay on the safe side. But what if a software failure happens? Or, even worse, the entire network self-driving cars share gets hacked and an outsider will be able to drive vehicles remotely?

Are driverless vehicles safe?

That brings us to the vast area of security concerns related to the cars that drive themselves. Visions of multiple crashes and situations when you’re sitting in a car that’s about to hit a pedestrian or a tree – and there’s nothing you can do – may certainly prevent many people from getting in the self-driving car.

Despite the huge trust engineers place in algorithms standing behind autonomous vehicles, you have to keep in mind that every software fails from time to time. Why should systems that are responsible for autonomous driving be an exception to this rule? In their case, however, there may be no time to wait for a bugfix to take place...

However, many experts claim that self-driving cars may be even safer than those driven by people. They can react faster than humans, don’t make human errors, are able to monitor many areas at once, and, most importantly, can anticipate what will happen next and act accordingly, which people are incapable of.

There are many life-changing, advanced safety features autonomous vehicles may use, including automatic emergency braking (AEB), blind spot warning (BSW), forward collision warning (FCW), rearview video system (RVS), or lane keep assist (LKS) to make AV’s use safer.

The safety potential is, then, significant, but to say more about the reality, we need to wait and see what the future will bring. For sure, more needs to be done to enhance each aspect of driverless vehicles' security, because a major issue or media-covered crash can put people off from the AV idea for years.

The future of self-driving cars

The future of autonomous cars is the great unknown, despite some significant developments publicized regularly. For example, self-driving taxis are already here, as ones deployed by Amazon hit the California public streets in February 2023. The company was testing its fleet of robotaxis with employees used as passengers. Waymo One, an autonomous ride-hailing service, already works in Phoenix and San Francisco. 

We may expect self-driving development to accelerate as many technological giants have shown interest in this area. Among the companies which have already invested in or cooperate with autonomous vehicle manufacturers, there are corporations like General Motors, Audi, BMW, Ford, Volkswagen, Volvo, Toyota, Honda, Bridgestone, and Microsoft.

Of course, there is no guarantee that an unknown, unpredicted obstacle won’t come their way. For example, the 2019 manslaughter case, with a Tesla vehicle driving with the Autopilot feature turned on involved, had far-reaching consequences, with regulators investigating this and similar crashes carefully.

Actually, the obstacle may also be a well-known one, namely: a hacker threat. The fear of hackers taking over the AV network has been paralyzing some of the market decision-makers for years. In fact, some experts say that the fear of hackers will stop the self-driving cars idea from spreading to a wide scale.

Driverless cars – key takeaways

Driverless cars seem to be the future of humanity, shaking up hopes about fewer car accidents, smoother traffic, and greater comfort of life – provided everything goes fine. As of 2023, the fast track to success seems to be probable, but it’s not the only option, and we don’t know exactly what’s in the cards for autonomous driving. It may be a bumpy ride for AV, but AI’s power will certainly give it a boost.

What may stop or slow down automated driving’s development are security issues. In the case of autonomous vehicles, the risk of an accident can’t be reduced to zero, but it’s also not zero for human drivers and regular cars. In fact, the human factor – or human error – in the form of distraction, intoxication, or fatigue is a major cause of car accidents, responsible for over 90% of them.

Software development services by Codete >

Excited about AI-based inventions and enhancements? Want to know how to use them best to your organization’s advantage? If you need advice on how Artificial Intelligence can help your company fly, don’t hesitate to contact Codete now.

Rated: 5.0 / 1 opinions
Avatar Female 7c9e57b38c

Dominika Reszke

IT Content Writer with 12 years of professional writing experience. Prefers facts and figures to any kind of fiction.

avatar male f667854eaa

Piotr Wawryka

Piotr has over 5 years of commercial experience writing Python applications. He is a software developer and data scientist at Codete since 2017 and a Ph.D. student at AGH University of Science Technology. His main field of interest is Neural Networks and their practical applications. He gives speeches at meetups and international conferences.

Our mission is to accelerate your growth through technology

Contact us

Codete Global
Spółka z ograniczoną odpowiedzialnością

Na Zjeździe 11
30-527 Kraków

NIP (VAT-ID): PL6762460401
REGON: 122745429
KRS: 0000983688

Get in Touch
  • icon facebook
  • icon linkedin
  • icon twitter
  • icon instagram
  • icon youtube
  • icon github
  • Kraków

    Na Zjeździe 11
    30-527 Kraków

  • Lublin

    Wojciechowska 7E
    20-704 Lublin

  • Berlin

    Bouchéstraße 12
    12435 Berlin