ADA TURING TEST

Artificial intelligence will be our chauffeur, co-driver, guardian & instructor but what should our minimal performance expectation be?

Public trust in Artificial Intelligence (AI) is essential if we are to realise its potential to reduce the annual 1.3 million road fatalities and 50 million injuries. Trust in the technology. Trust in the brands offering enhanced mobility services. Yet most importantly, trust in the driving behavior of AI systems deployed on our roads.

Trust is the foundation of safety within our global road network. Trust that others will compensate for your mistakes and will respect our limited physical abilities to tolerate crash forces. Trust in the shared social responsibility to act with care to prevent crashes, provide post-crash care and collaborate to mitigate individual failures.

To establish this level of trust the ADA Turing Test proposes a minimum performance threshold expected for AI systems deployed on our roads. The rules apply equally to AI for advanced driver assistance systems (ADAS) or within Self-Driving Systems (SDS), Automated Driving Systems (ADS) or Automated Vehicles (AV).

The Three Laws for AI on our Roads

1. Prove AI meets, or exceeds, the performance of a competent & careful human driver

2. Prove AI never engages in careless, dangerous or reckless driving behaviour

3. Prove AI remains aware, willing and able to avoid collisions at all times

Responding to this critical issue, the ITU (International Telecommunications Union) and ADA have formed a new​ joint initiative to organise the first #AIforGood road safety event hosted at ITU Telecom World in Budapest on Tuesday, 10 September 2019. The objective of this ITU/ADA Workshop will be to present the state of the art in autonomous driving solutions and work together to define the “ADA AV Turing Test” for AI on our roads.​

For more information please visit our events page.