The government is to begin introducing artificial intelligence into the civil service to “support civil servants in their work”, Research Deputy Minister Nicodemos Damianou said after Wednesday’s cabinet meeting.

He described the move as “an important step in the process of digital modernisation of the public sector” and “part of the broader effort to utilise the capabilities of artificial intelligence in the public sector”.

As for how it will help civil servants, he said artificial intelligence will both facilitate and accelerate the completion of tasks and “allow them to focus on what has real value”.

“By automating simple actions, rapidly analysing large amounts of information, and integrating security mechanisms, this artificial intelligence tool enables us to work smarter and save time with complete protection of personal data and public information,” he said.

“The government will now be making use of generative artificial intelligence tools in the public sector in a structured way, so as to ensure the security of the data being circulated.”

He added that the first stage of artificial intelligence’s implementation will be the granting of 350 licences for Microsoft’s Co-Pilot tool, with civil servants set to be offered training programmes to ensure that the technology will be “used safely, responsibly, and with a substantial impact on daily work”.

He said training on how to use Co-Pilot will be offered by Microsoft, and that “practically everyone” in the civil service will “now have these tools built into their computer or device connected to Microsoft 365 to use in their everyday lives”.

Meanwhile, he said, the government is also moving forward with a “series of actions for the broader adoption and utilisation of artificial intelligence in the public sector”, for which it will invest an initial €5 million.

Businesses, he said, are “invited to develop solutions based on artificial intelligence for the state’s real challenges, from the interconnection of education with the labour market to the [prediction] of extreme weather phenomena”.

Later on Friday, the Cyprus Mail spoke to the deputy ministry’s spokeswoman Nasia Zanti, who stressed that decision-making responsibility will still rest with “the human factor”, and that any civil servant who makes use of artificial intelligence will be required to check the results of the end product themselves.

“We will explain to them very clearly exactly how to use it and the boundaries which will be set,” she said.

She also gave the example of speechwriting as one aspect of government work which could be aided by artificial intelligence, explaining that one could find the information about which they wanted to write a speech and have artificial intelligence write them a draft.

But once it is written, she stressed, a human would look over it.

She was also keen to state that the data shared with Co-Pilot by civil servants “will not be shared anywhere” and will be kept in a “closed government environment”, from which data will not leak or be shared.

The announcement comes as questions continue to be raised abroad over the use of artificial intelligence in official contexts.

In Australia, accounting firm Deloitte announced on Monday that it will partially refund payment for an Australian government which contained multiple errors after being partly produced by artificial intelligence.

Australia’s employment department had commissioned a AU$439,000 (€248,000) review to help assess problems with the country’s welfare system, with the Australian financial review finding that the document contained multiple errors, including references to and citations of non-existent reports written by academics.

Newspaper the Financial Times described the incident as an “embarrassing episode” which “underscores the dangers posed … by using AI technology, particularly the danger of ‘hallucinations’”.

Meanwhile, research from the Massachusetts Institute of Technology found that 95 per cent of organisations are receiving zero return from their investments in generative artificial intelligence, with economic experts now speaking of an “AI bubble” which could “burst” in the near future.

This was corroborated by the Bank of England’s financial policy committee on Thursday, which said there is a growing risk of a “sharp market correction” as investors overvalue the potential of artificial intelligence.

“On a number of measures, equity market valuations appear stretched, particularly for technology companies focused on artificial intelligence. This … leaves equity markets particularly exposed should expectations around the impact of AI become less optimistic,” it said.

As such, it warned that a “sudden correction could occur”, which would, in essence, see share prices tumble and have knock-on impacts for the wider economy.