When the digital avatar “Diella” was launched by the Albanian government last year, its purpose was clear: to serve as an incorruptible public servant in a country known for widespread corruption, according to new reporting by The New York Times. 

In practice, Albania’s so-called “AI minister” has already effectively achieved that end, taking on the role of the digital official that can’t be bribed for quicker service whenever citizens apply for government services online. Soon, Diella will also be able to collate applications for governmental contracts to impartially assess and identify qualified bidders, the newspaper said. Diella can also be audited, the Times noted.

“I have no personal ambitions or interests,” Diella said in a video aired last September that was cited in the report. “I have only data, knowledge and algorithms dedicated to serving citizens impartially, with transparency, and without ever growing tired.”

Yet the program’s reputation is not entirely untarnished. 

Leaders of Albania’s National Information Agency, the government body that built the avatar and runs the state’s digital infrastructure, are now under investigation in a case involving bid-rigging and intimidation linked to contract applications, the Times reported.

Last month, prosecutors tied the agency’s director and her deputy to a criminal syndicate accused of using intimidation to manipulate contract applications. The pair have been placed under house arrest but not formally charged, according to the report. 

Whether Delia will be an effective anti-corruption tool for Albania remains to be seen. In December, experts at the German Council on Foreign Relations concluded that the AI program has improved “efficiency and impartiality of public procurement decisions taken by the government,” but the country’s record of meeting EU standards remains “problematic,” according to the report. 

The allegations against Delia’s creators raises questions about whether the AI minister can be programmed to overlook graft, or else be fed information that leads to inaccurate conclusions, The Times reported. 

Read more at The New York Times