Armilla Review #79
TOP STORY
The CFO's Playbook for Navigating Generative AI Risks
As businesses rapidly adopt generative AI for tasks like document analysis and memo writing, CFOs find themselves at a crossroads between embracing innovation and mitigating risks. Generative AI, while powerful, often behaves unpredictably, leading to bizarre errors and potential disruptions in workflows. Instances like ChatGPT's nonsensical responses highlight the technology's immaturity and the dangers of "hallucinations" that could harm a company's reputation. Finance leaders worldwide are weighing the benefits against risks such as data security breaches, operational errors, and financial over investment. They emphasize the importance of setting clear policies, training employees, and ensuring human oversight to navigate this evolving technological landscape effectively.
Source: Financial Management Magazine
?? Sign up for our newsletter now to get the latest updates in your inbox.
THE HEADLINES
Uncertain Futures: Why Policymakers Must Act on AI Risks
The uncertain yet potentially catastrophic risks of advanced AI are prompting urgent calls for regulation, despite the difficulty in precisely quantifying these dangers. Critics argue that without exact probabilities, policymakers should refrain from regulating AI technologies. However, history shows that action under uncertainty is not only common but necessary, especially when potential harms are immense. By adopting precautionary regulatory approaches, society can mitigate risks even when they are not fully understood. The article emphasizes that waiting for precise calculations may lead to inaction in the face of existential threats, advocating for immediate policy interventions based on expert estimations and historical parallels.
Source: LAWFARE
LinkedIn's Silent Data Harvesting: AI Training Without Consent?
LinkedIn has come under scrutiny for using user data to train AI models without adequately updating its terms of service or informing users beforehand. U.S. users discovered an opt-out option buried in settings, but the company's initial failure to transparently communicate this practice has raised privacy concerns. While LinkedIn claims the data aids in improving features like writing suggestions, critics argue that users were not given proper notice or the opportunity to consent. Privacy advocates and regulatory bodies are calling for investigations, highlighting the tension between AI development and user privacy rights.
Source: TechCrunch
领英推荐
Big Tech's Emission Gap: The Truth Behind Data Center Pollution
A recent analysis reveals that the emissions from in-house data centers of major tech companies like Google, Microsoft, Meta, and Apple are significantly higher than officially reported—potentially over seven times higher. These companies often use renewable energy certificates to offset their emissions on paper, creating a misleading picture of their actual environmental impact. As the demand for data processing grows, especially with the rise of AI technologies, energy consumption and associated emissions are expected to increase dramatically. Experts and environmental groups are calling for more transparent and accurate reporting methods to ensure accountability and drive meaningful action toward reducing carbon footprints.
Source: The Guardian
Microsoft's AI Paradox: Climate Commitments vs. Fossil Fuel Deals
Despite publicly championing AI's potential to combat climate change, Microsoft is simultaneously marketing its AI technologies to fossil fuel giants like ExxonMobil and Chevron to enhance oil and gas production. Internal documents and employee accounts suggest a dissonance between Microsoft's environmental commitments and its business practices. A whistleblower complaint alleges that the company failed to disclose the environmental harms associated with these technologies in its public reports. This revelation raises questions about the ethical responsibilities of tech companies and the genuine impact of their sustainability pledges.?
Source: The Atlantic
Harvesting Without Consent: Meta Admits Using Aussie Photos for AI
Meta has admitted to using photos and posts shared by Australians on Facebook and Instagram to train its AI models without offering an opt-out option available in other countries. During a Senate inquiry, Meta executives revealed that while they exclude content from users under 18, images of children posted by adults are still used. This disclosure has prompted Australian politicians and academics to demand stronger legal protections against unauthorized data harvesting. The incident underscores the need for updated privacy laws to keep pace with technological advancements and to safeguard personal information from being used without explicit consent.
Source: City News
PEOPLE & AI
In our latest episode of People & AI, we had the privilege of speaking with Benjamin Roome, PhD, founder of Ethical Resolve and CEO of Badge List, about the pivotal intersection of AI, ethics, and innovation. Key takeaways: > Disparate Impact & Legal Standards: Explore the relevance of thresholds like the four-fifths rule and why ethical AI requires surpassing legal compliance to ensure fairness. > Ethical AI Governance: Learn the importance of establishing clear ethical AI policies and governance frameworks to foster accountability and transparency. > Risk Mitigation: Understand the role of ethical design, stakeholder impact assessments, and addressing broader social and economic risks. Listen to the full episode to learn how AI can amplify human capacity, integrate ethics into business processes, and navigate the complexities of AI regulation.
Apple podcasts: https://lnkd.in/ga4t4WuZ
Spotify: https://lnkd.in/gBzmKsDE
YouTube: https://lnkd.in/gFWGHRHy