- By Zoe Kleinman, Philippa Wain and Ashleigh Swan
- Technology team
Discrimination is a more urgent concern for the advancement of artificial intelligence than for human extinction, says the EU competition chief.
Margrethe Vestager told the BBC that ‘guardrails’ are needed to counter the increased risks of technology.
He said this has been key where AI is used to help make decisions that can affect someone’s livelihood, such as whether they can apply for a mortgage.
The European Parliament will vote on its proposed AI rules on Wednesday.
The AI Act is being heeded by politicians amid warnings about the development of technology, which allows computers to perform tasks that typically require human intelligence, too quickly.
In an exclusive interview with the BBC, Vestager said the potential for AI to amplify bias or discrimination, which can be contained in the vast amounts of data coming from the internet and used to train models and tools, was a more pressing concern. .
“Probably [the risk of extinction] may exist, but I think the probability is quite small. I think the risks of AI are more that people will be discriminated against [against]they will not be seen for what they are.
“If it’s a bank that uses it to decide whether or not I can get a mortgage, or if it’s your council’s social services, then you want to make sure you’re not being discriminated against [against] because of your gender or your color or your zip code,” she said.
On Tuesday, the Irish data protection authority said it had suspended Google’s planned EU rollout of its AI chatbot Bard.
It said it had been notified by Google that its competitor ChatGPT would be introduced in the EU this week, but had not yet received any details or information showing how the company had identified and minimized data protection risks for potential users.
Deputy Commissioner Graham Doyle said the DPC was seeking the information “urgently” and had raised further data protection requests to Google.
Ms Vestager, who is the executive vice president of the European Commission, said AI regulation must be a “global affair”.
He insisted on consensus among “like-minded” countries before involving more jurisdictions, such as China.
“We are starting to work on a United Nations approach. But we shouldn’t hold our breath,” he said.
“We should do what we can here and now.”
Ms Vestager is leading EU efforts to create a voluntary code of conduct with the US government, which would see companies that use or develop AI adhere to a set of standards that are not legally binding.
The current draft of the AI Act seeks to categorize applications of AI into levels of risk to consumers, with AI-enabled video games or spam filters falling into the lowest risk category.
High-risk AI systems include those used to evaluate credit scores or access to loans and housing. This is where the strict controls on the technology will focus.
But as AI continues to develop rapidly, Vestager said you need to be pragmatic when it comes to fine-tuning the rules around this technology.
“It’s better to get, say 80% now than 100% ever, so we start and then come back when we learn and then correct with others,” she said.
Ms Vestager said there was “definitely a risk” that AI could be used to influence the upcoming election.
He said the challenge for police and intelligence services would be to be “completely at the top” of a criminal sector where there is a risk they will get ahead in the race to use the technology.
“If your social feed can be scanned to get a complete profile of you, the risk of being manipulated is just huge,” he said, “and if we end up in a situation where we don’t believe in anything, then we’ve completely undermined our society .”
But Ms Vestager said this was unrealistic.
“Nobody can enforce that. Nobody can make sure everyone is on board,” he said, noting that a break could be used by some as an opportunity to get ahead of competitors.
“What I think is important is that every developer knows that everyone has signed up for the same guardrails so that no one takes too much risk.”
The European Parliament’s proposals for the AI law aim to limit the use of biometric identification systems and the indiscriminate collection of user data from social media or CCTV footage for purposes such as facial recognition systems.
However, Vestager said: “We want to put in strict guardrails so that it’s not used in real time, but only in specific circumstances where you’re looking for a missing child or there’s a terrorist on the run.
“Parliament has a much more principled position that it will vote on tomorrow to essentially ban it completely.”
Before the AI law can be finalized as the world’s first regulation on the use and development of AI systems, the three powers of the EU: the Commission, the Parliament and the Council will have to agree on its final version.
It is not expected to come into force before 2025.
#Warning #loans #mortgages #big #risk #BBC #News