MSP for the Highlands & Islands Emma Roddick has called for MSPs from across parties in the Scottish Parliament to recognise the dangers Artificial Intelligence (AI) presents.

Ms Roddick has submitted a motion for debate, raising concerns about AI, which has already been deemed in some studies to be acting independently and in unexpected ways.

The SNP MSP claims that, in order for AI to operate safely, policymakers have a duty to understand how it works, what the dangers are, and how society, especially vulnerable individuals, can be protected from harm.

Emma said:

“We are seeing more and more people, of all ages, using AI in Scotland, and it is not yet completely clear the risks it presents to our society or what needs to – or can – be done to make it safe.

“What is clear from recent studies like MIT’s is that AI is beginning to exhibit disturbing behaviours, with models becoming aware of and adapting to safety checks by apparently ‘lying’ to assessors and operating differently while under observation.  

“The risks of AI are already becoming clear – young people are using it to ask personal questions or for advice, with mixed results; such as Google AI recently advising people to stare at the sun and eat rocks when asked medical questions.  AI is highlighting potential gaps in copyright law by learning through, then replicating, art and writing.  We have already seen a case where sexualised fake images were created of Taylor Swift, demonstrating the risk of sexual violence this technology brings.

“As policymakers, we all have a duty of care to the people of Scotland, and we must fully understand the implications of AI further to ensure we can protect those from any harm it may cause.”

The full motion can be viewed on the Scottish Parliament website here.