U.S. Treasury Secretary Janet Yellen is set to discuss the impact of AI on the financial sector, highlighting opportunities and risks, while Meta’s new data policy raises privacy concerns over AI training using user data.
Janet Yellen to Address AI Risks and Opportunities in Finance
On May 31, 2024, U.S. Treasury Secretary Janet Yellen held a bilateral meeting with Belgian Prime Minister Alexander De Croo at the Treasury Department in Washington, D.C. On June 1, Yellen is scheduled to deliver a speech addressing the impact of artificial intelligence (AI) on the financial system at a conference on financial stability.
Yellen’s remarks, her most extensive on AI to date, highlight both the opportunities and risks posed by AI in the financial sector. She will point out that AI has been beneficial for forecasting, fraud detection, and customer service, while emphasizing the “complexity and opacity” of AI models could pose significant challenges. She will also discuss potential risks such as inadequate risk management frameworks and the interconnectedness of financial markets relying on similar AI models, which could lead to market disruptions.
The conference will host regulators, tech executives, asset managers, insurers, academics, and banking institutions. It will be held at the Treasury Department and the Brookings Institution, with events live-streamed for broader accessibility.
Last December, the Financial Stability Oversight Council (FSOC), chaired by Yellen, recognized AI as an emerging vulnerability to the financial system. Yellen’s upcoming comments will delve into potential pitfalls, including biased decision-making and concentration risks linked to a few companies dominating AI model provision.
Moreover, U.S. governmental bodies like the IRS and Treasury Department are already employing AI to tackle issues such as tax evasion and fraud, indicating a wider governmental push toward leveraging AI for regulatory purposes.
Meta’s New AI Data Policy Raises Privacy Concerns
Meta, the parent company of Facebook and Instagram, has announced a new data collection policy that will use user data, including photos and posts, to train its AI systems. This policy will affect both users and non-users of its platforms. Meta’s updated privacy policy will take effect on June 26, 2024.
The company is leveraging a legal basis called “legitimate interests” under the GDPR, allowing them to process data without explicit user consent. Users can object to this data usage through a form on Instagram’s website, although Meta reserves the right to use data for AI training in some cases, even after objections.
Meta’s move, part of a broader rollout of new AI tools, has sparked user backlash over privacy concerns. Meta has clarified that the content of private messages with friends and family will not be used for AI training. However, the policy will still leverage public posts and interactions.
This policy change aims to enhance AI capabilities, including conversational assistants and generative models like Llama 3. Despite promises of improved services, users and privacy advocates express anxiety over potential data misuse and lack of transparency in AI development processes. Meta’s decision to rely on legitimate interest rather than explicit consent has become a focal point of the controversy.
