Skip to main content

Governing transport in the algorithmic age

Transport policy needs to ready itself for the age of algorithms, and policy makers must become algorithmically literate. This is the key message of a new report by the International Transport Forum, presented today (23 May) at the global summit of transport ministers in Leipzig, Germany.

Automated decision-making is becoming more and more prevalent. Choices that used to be made by humans are instead entrusted to algorithms and based on Artificial Intelligence (AI). Transport is one of the areas where algorithms play an increasing role, for instance in automated driving or new mobility services.

Algorithms can be hugely beneficial. They are able to solve formerly intractable problems or improve our ability to accomplish previously time-consuming tasks.

They also raise unique legal, regulatory and ethical challenges. Algorithmic decisions may result in unintended and harmful behaviour – for instance where the wrong objective is specified, or if the training data for machine learning is biased or corrupted. When algorithms fail, people can get hurt and material damaged. Where these risks can propagate across systems, the harm can multiply.

Privacy risks also exist. Algorithms are data-processing technologies. Yet data anonymisation is rarely robust enough to stand up against serious data-discovery attacks. Vulnerabilities grow as adversarial algorithms get better at extracting data.

Physical and moral hazards emerge particularly when AI systems start to drift into areas of human decision-making in ways that remain inscrutable to human cognisance.

Algorithmic systems are highly opaque and difficult to explain to regulators, or to those affected by their decisions. Code is often created in environments that are not open to scrutiny, such as private companies. It uses machine languages that are not widely understood. The operation of several types of AI algorithms may not even be explained by their designers.

The lack of insight into AI processes challenges traditional forms of public governance. Transport policy, its institutions and regulatory approaches have been designed for human decisions. They are bound by legal and analogue logic that is now challenged by systems which function with machine logic.

Public authorities therefore have to evaluate whether their institutions and working methods are adapted to this development. If not, they will have to begin to reshape themselves for a more algorithmic world. This will require new skill sets, notably code literacy.

The report “Governing Transport in the Algorithmic Age” makes a number of further, specific recommendations. Among other things, it suggests that public authorities:

  • Convert analogue regulations into machine-readable code - for example, authorities could encode permissible uses of street and curb-space as the Los Angeles’ open-source Mobility Data Specification (MDS) does.
  • Use algorithmic systems to regulate more dynamically and efficiently - AI may create new ways of regulating with a lighter touch.
  • Compare the performance of algorithms with that of humans - is the balance of risks and benefits tilted towards one or the other?
  • Establish robust regulatory frameworks that ensure accountability for decisions taken by algorithms - ensure that algorithmic systems are built so they can be trusted.
  • Establish clear guidelines and regulatory action to assess the impact of algorithmic decision-making - such as Canada’s “Directive on Automated Decision-Making”, a model approach.

The work for this report was carried out in the context of a project initiated and funded by the International Transport Forum's Corporate Partnership Board (CPB). The Corporate Partnership Board (CPB) is the ITF’s platform for engaging with the private sector and enriching global transport policy discussion with a business perspective.

The findings are those of the involved parties; they do not necessarily reflect the views of ITF member countries. The CPB companies involved in this project are: Abertis, ExxonMobil, Kapsch TrafficCom, Latvian Railways, NXP, PTV Group, RATP Group, Renault Nissan Mitsubishi Alliance, Robert Bosch GmbH, SAS Institute, Siemens, SNCF, Total, Toyota Motor Corporation, Uber, and Valeo. 

Download the report: https://www.itf-oecd.org/governing-transport-algorithmic-age

Also just publised:

Expanding Innovation Horizons: Learning from Transport Solutions in the Global South
https://www.itf-oecd.org/expanding-innovation-horizons-learning-transport-solutions-global-south

New Directions for Data-driven Transport Safety
https://www.itf-oecd.org/new-directions-data-driven-transport-safety-0

Media Contact:
Michael KLOTH
Head of Communications
M +33 (0)6 15 95 03 27
michael.kloth@itf-oecd.org

Summit media resources: