Microsoft CPO highlights opportunities and challenges of AI
Whilst AI brings risks and challenges, it also promises to have a transformative positive impact across numerous areas, Julie Brill, Microsoft’s Chief Privacy Officer and Corporate Vice President Global Privacy, Safety, and Regulatory Affairs told delegates at CPDP (Computers, Privacy and Data Protection) conference.
“Often these conversations immediately focus on risks and leave out the opportunities. I think it’s really hard to develop appropriate balanced regulation if you’re not thinking about the full landscape”, she said.
Brill, who was formerly Commissioner at the US Federal Trade Commission, drew attention to some of the challenges of AI, in particular, the risk of disinformation and bias.
“There are issues with disinformation, scams, and a whole bunch of other issues which are real and definitely need to be addressed and we are probably going to need to pay even more attention to some of those issues as AI continues to develop.
“But these are issues we already have. Sadly, we have discrimination and bias in the world. We also have laws that address them but we’re going to have to pay even more attention as AI rolls out more, so I am very, very optimistic.”
She detailed some of the specific benefits AI can bring, such as addressing global inequalities in healthcare in the context of eye disease, highlighting the lower proportion of ophthalmologists in the Global South.
“Individuals have serious issues that would cause them to go blind if untreated. Through AI there is a way to detect this in a much more simple manner and address these diseases with early interventions that might not require ophthalmologists in all cases.”
The democratization of AI
Brill also highlighted the “democratization of AI” through the use of productivity tools in workplace settings explaining the improvement in efficiency that can be offered through these solutions.
“You all know Word or Google Docs. You write a memo and then you want to turn that memo into a PowerPoint. You give that to an assistant or you would sit down and do that, and that would take a while to do.
“With a couple of buttons in the context of generative AI, you can immediately take the memo and turn it into a PowerPoint, and if you want to adjust it you can adjust it.”
A major challenge to harnessing the opportunities of AI is data access.
“Data is going to be an incredibly important element as we build AI,” said Brill. “It will be very hard for EU to accelerate at the same level as the US. The idea of building AI that is appropriately calibrated for Europe is deeply important.
“You need land you can build data centres on; you need these to be carefully monitored; you need CPUS with the right type of chips. The level of investment we’re making in Europe to make that real is astronomical.”
Could the AI Act become a global standard like GDPR?
The EU is the first jurisdiction in the world to apply AI-specific legislation, through its AI Act. RAID Director Ben Avison asked Brill if she thought this risk-based approach could become a global standard like GDPR.
“GDPR became exportable because every jurisdiction wanted to interact with Europe,” she said. “In the AI Act there is no requirement that other jurisdictions take the same approach, but we’re seeing a deep convergence. A consensus is being built around the fundamentals of what we want AI to be – there is a strong theme of taking a risk-based approach. The technology needs to be trusted and embraced and the way it’s going to happen is by exporting this risk-based approach.”
Brill was joined on the panel by Anna Buchta, Head of Policy & Consultation Unit at the European Data Protection Supervisor (EDPS), who said “There is consensus internationally that the risk-based approach is the right framework. But some jurisdictions would not go near a ban on even the highest risk applications.”
Both Brill and Buchta spoke at RAID 2023, which also centred on the theme of the complex balance policymakers must strike when regulating transformative technology.
RAID 2024, in Brussels on 23-24 September, includes a panel “Grasping the Opportunities and Threats of AI and Digital Transformation”.
Confirmed RAID 2024 speakers include: Didier Reynders, Commissioner for Justice, European Commission; Anu Talus, Chair, European Data Protection Board (EDPB); Katherine Harman-Stokes, Acting Director, Office of Privacy and Civil Liberties, US Department of Justice; Dr. Leonardo Cervera Navas, Secretary General, European Data Protection Supervisor (EDPS); Jon Donenberg, Deputy Director – Technology, National Economic Council, USA; John Edwards, UK Information Commissioner, Information Commissioner’s Office; Brando Benifei MEP and many more leading policymakers and regulators.
