How government use of AI could harm democracy

Many countries are investigating how artificial intelligence can help with everything from treatment taxes to determining welfare services. But a study shows that citizens are not as enthusiastic as their governments – and this can create real risks to democracy.

“Focuses only on short -term efficiency gains and shiny technology risks that trigger public setbacks and contribute to a long -term decline in democratic trust and legitimacy,” says Alexander Wuttke at Ludwig Maximilian University of Munich in Germany.

Wuttke and his colleagues asked about 1200 people in the UK to share their feelings about government acts, where Eith a human or an AI handled the task. These hypothetical scenarios included the treatment of tax returns, appning or rejection of welfare applications and made risk assessments as to whether the defendant should be entitled to bail.

Some people were told only about how AI could improve government efficiency-but others learned about both AI-related benefits and risks. These risks include difficult to understand how AI decisions are made, growing government addiction that it becomes reversible over time and lack of clear ways for citizens to compete and correct AI decisions.

As people became aware of AI-related risks, they report both a marked decrease in their confidence in the government and a feeling of losing control. For example, the percentage of participants who reported loss of democratic control in their government rose from 45 percent to more than 81 percent in scenarios, with the government becoming more and more dependent on AI for dealing with specific tasks.

The proportion of people who demanded less AI in the government also rose sharply when participants learned about the risk – rising from less than 20 percent in the baseline scenario to more than in any scenario where they learned about both the benefits and risks of the government using AI.

Despite these results, Democratic Governments could make use of AI responsible for retaining citizens’ confidence, says Hannah Quay-De La Vallee at the Center for Democracy & Technology in Washington DC. But she says there are successful stories about AI in government so far. Meanwhile, there are already “quite a few failure – and the efforts of these boxes can be incredibly high.

For example, US state efforts to automate the treatment of public benefits demands that tens of thousands of people are incorrectly accused of fraud. Some of these people ended up having to fil for bankruptcy or lose their homes. “Government errors have huge, long-range effects,” says Quay-De La Vallee.