‘AI is getting better at a range of tasks, including many areas thought to be the province of human beings’
– World Economic Forum
The World Economic Forum says the impacts of AI on journalism are similar in many areas:
- automating routine operations
- gaining faster insights
- lowering barriers to entry
They identify seven challenges for journalism. Some are internal issues, such as verifying authenticity and the possible impacts on copyright and fair use. Others reflect changes in the news environment, for example, the growing prominence of data as an essential ingredient
Here is their list:
- Availability of data – Machine learning systems require very large amounts of data.
- Understanding unstructured data – Systems work best with organized datasets, for example, those already in a database. Data that is insufficiently tagged is difficult to process.
- Having self-awareness – Algorithms cannot explain how an outcome was reached. This creates difficulties for verification and accountability. Disclosure is a related issue: should AI involvement in a story be publicly identified?
- Verifying authenticity – Systems operate with any data presented. Until new measures are developed, systems cannot differentiate what is ‘good data’.
- Redefining copyright and fair use – Machine learning results come from works created by humans. This prompts legal questions about ownership of works derived by AI systems.
- Ensuring corporate accountability – The WEF report presumes legal and moral accountability resides in production owners.
- Exacerbating asymmetrical power – The build-up of AI capabilities will favour larger organizations. Concentrations could result in big news organizations, or even in big tech companies.
Their study presents useful exhibits for:
(a) A timeline of AI competing with humans
(b) The AI value chain in journalism