Welcome to Your New Car! Please Set Your Morality Preferences..

Imagine as you step into your newly purchased Level 5 Autonomous Vehicle - a sleek, AI-powered machine designed to drive itself under all conditions.

The seats adjust automatically to your posture, the dashboard illuminates, and a futuristic assistant materializes before you.

"Welcome! Before we begin, let's configure your ethical preferences."

1. Uncovering the Ethical Questionnaire!

You blink in surprise. Ethical preferences? You were expecting a Wi-Fi setup or seat heating controls, not a moral questionnaire.

The assistant continues -

• “In case of an unavoidable crash, should the vehicle prioritize your life over pedestrians?”
• “Would you like your vehicle to strictly follow all traffic laws, or should it have discretion in life-threatening situations?”
• “If given a choice between swerving into a single cyclist or hitting a stationary object, which would you prefer?”

You hesitate. You thought the hardest part of buying the car was picking the right color and selecting the most needed features, not deciding on life-and-death rules.

The assistant prompts the next set of questions -

• “If you are experiencing a medical emergency and need to get to a hospital, should your vehicle exceed speed limits and bypass traffic regulations in a safe and controlled manner?”
• “If any legal issues arise from an incident while driving in autonomous mode, do you consent to disclose vehicle data as evidence in court?”

You take a deep breath. The implications of these choices weigh heavily on you. The assistant reassures you:

"These settings will shape how your vehicle behaves in critical situations. Would you like to proceed with the standard ethical framework, or customize your preferences?"

You hesitate again. Do you trust AI to make the right choices, or do you want control? But first, let’s take a moment to reflect - when learning to drive, didn’t we already consider these dilemmas?

After all, when obtaining a driving license or stepping into our very first car learning session, we encountered similar situations.?Then why do these questions feel strange when a vehicle asks us?

Let’s dive deeper to understand how human driving ethics compare to those of an autonomous vehicle in the future.

2. 'Seeing the road' with Human Instinct vs. AI Logic

Driving is more than just following traffic laws. Humans instinctively assess multiple layers of information simultaneously - legal rules, social norms, environmental conditions, and ethical considerations - to make decisions on the road.

These factors influence every maneuver, from changing lanes to reacting to sudden obstacles.

2.1 What Humans Consider While Driving?

1. Traffic Laws and Road Signs – Speed limits, stop signs, lane markings, and traffic lights set the foundation for driving behavior, but humans sometimes make judgment calls (e.g., cautiously running a red light to allow an ambulance to pass).
2. Predicting Intentions of Others – Unlike AI, humans make eye contact with pedestrians, notice hesitant cyclists, and interpret body language to predict movements.
3. Social Driving Norms – In different cultures, honking may be considered rude or essential communication. Merging into traffic in one country might involve strict lane discipline, while in another, it relies on subtle negotiations between drivers.
4. Weather and Road Conditions – Humans adjust their driving based on rain, snow, or low visibility, sometimes disregarding official rules to stay safe.
5. Situational Ethics and Quick Decisions – A driver might break a minor traffic rule if it prevents an accident, such as performing a rapid lane change to avoid a collision.

2.2 How Autonomous Vehicles See the Road Differently?

Autonomous vehicles process information purely through data, sensors, and predefined decision-making models rather than intuition or social interactions.

1. Sensors and Cameras: Detect nearby vehicles, lane boundaries, traffic lights, and obstacles.
2. Artificial Intelligence Models: Calculate probabilities for the safest actions based on historical data and real-time road conditions.
3. Legal Compliance: AVs are programmed to strictly follow traffic rules, but edge cases - such as deciding between hitting an animal or performing an evasive maneuver - require additional "ethical programming".
4. Standardized Reactions: AVs do not predict intent the way humans do. Instead, they react based on hardcoded rules and statistical models, which might not account for human subtleties like eye contact.

2.3 Now, The Fundamental Difference!

Humans rely on intuition, learned behavior, and social interactions to drive, whereas AVs rely on sensor input, probability calculations, and predefined rules.

The result is that while AVs are consistent and methodical, they may struggle in ambiguous or unpredictable situations where humans rely on instincts.

With these differences in mind, we must ask:

Will AVs ever be able to replicate the full spectrum of human driving judgment, or will they forever remain constrained by programmed logic?

3. The 'Trolley Problem' and 'Probability Paradox'

The Trolley Problem is a classic ethical dilemma that presents a scenario in which a person must choose between saving one group of people or another. Traditionally, a runaway trolley is heading toward five people on the tracks.

The only way to save them is to pull a lever that diverts the trolley onto another track, where it will hit a single person instead. This forces a choice between minimizing casualties or avoiding direct responsibility for a death.

Autonomous vehicles face similar dilemmas when they must make split-second decisions in unavoidable accident scenarios.

However, unlike humans, AVs do not rely on instinct - they compute statistical probabilities to determine the best possible outcome.

3.1 How Humans React?

Humans make split-second decisions based on Gut instincts developed through years of experience, Quick moral judgment that is often inconsistent, Emotional bias, such as protecting family members first!

For example, a driver faced with an emergency might instinctively execute a sudden steering input, without calculating whether that is truly the best course of action.

3.2 How Autonomous Vehicles Typically Reacts?

Instead of instinct, AI relies on probabilities:

• Scenario A: The AV calculates a 90% chance that full braking will prevent an accident, versus a 10% risk of failing.
• Scenario B: A lane departure maneuver has a 70% chance of avoiding a collision but introduces a 30% risk of impacting another vehicle.
• Scenario C: Speeding up has a 40% chance of avoiding impact but a 60% probability of causing more harm.

The AI does not "choose" in the way a human does - it selects the action with the highest probability of minimizing damage.

3.3 Does This Lead to Better or Worse Outcomes?

• Positive Impact: AI eliminates panic-driven mistakes and ensures mathematically optimal decisions are made.
• Negative Impact: AI lacks human flexibility and instinct; in rare cases, a human may react in a way that defies probability but saves lives.

The Probability Paradox highlights a key dilemma: AI makes the best decision statistically, but does that always mean it’s the right one morally?

As we move toward a world of fully autonomous vehicles, we must decide: Should AI be trusted to make life-or-death calls based on probability alone?

4. When Two AVs Disagree in a Face-to-Face Event

4.1 The Collision Course Dilemma

Imagine two autonomous vehicles approaching an intersection from opposite directions. Both vehicles detect a potential collision scenario but have different decision-making algorithms.

One AV prioritizes minimizing harm to pedestrians, while the other is programmed for passenger self-preservation. Both vehicles anticipate the potential for impact, but they do not share the same ethical or risk-based decision models.

How AVs Compute Different Outcomes?

Each vehicle independently runs its onboard trajectory risk assessment and prediction algorithms - Vehicle A serves in Minimize Harm Mode, whereas Vehicle B serves in Self-Preservation Mode.

4.2 Ethical AI Conflict: Who Has the Right of Way?

Unlike human drivers, AVs do not negotiate intent through gestures or intuition. When two AVs with conflicting ethical programming meet, the result can lead to indecision, unexpected outcomes, or unintended escalation of risk.

Our Key questions -

• Should AVs share decision-making data in real-time to coordinate responses?
• Would a regulatory framework require all AVs to follow standardized ethical principles?
• What happens when manufacturers program different safety priorities into vehicles?

4.3 The Need for a 'Universal Ethical Traffic Regulator' Model

To prevent conflicting ethical decisions between AVs, future transportation systems may need:

• A central traffic arbitration system that dictates universal response behaviors?
• Vehicle-to-Vehicle (V2V) communication protocols to align decisions in real-time?
• Global Regulations enforcing a common safety philosophy for all manufacturers?

As AVs evolve, they must not only make individual ethical decisions but also cooperate in a shared mobility network. Without a universal standard, the risk remains that AVs will make competing ethical decisions, leading to outcomes that are difficult to predict or justify.

Will AVs ever reach a consensus on morality, or will conflicting AI decisions lead to a fragmented ethical future?

5. The Unfinished Code of Morality

Lets head back into our newly purchased Level 5 Autonomous vehicle!

As you configure your vehicle’s ethical preferences, you realize something unsettling - your answers define the morality of an AI system that will make life-or-death decisions on your behalf.

The assistant’s voice prompts one final question:

"Morality evolves. Humans learn from mistakes. Should I adapt over time, or should I always follow the choices you made today?"

You pause. The car is asking whether it should remain static in its ethics or evolve as it learns from the world around it.

What if AI eventually surpasses human morality, making better ethical choices than humans ever could? Or will AI always remain a reflection of human imperfection, frozen in ethical dilemmas we can never truly solve?

In the end, are we teaching AI morality, or is AI teaching us what morality truly means?

Further Reads

Should a self-driving car kill the baby or the grandma? Depends on where you’re from. | MIT Technology Review

Should self-driving Cars prioritize passenger safety or pedestrian safety in emergency situations? | by Fatiu O. Bello | Dec, 2024 | Medium

Trolley problem - Wikipedia