Why It's Dangerous For AI To Regulate Itself
In this Forbes Article Compliance.ai’s CEO and Co-Founder, Kayvan Alikhani, explains why it is imperative that companies and regulatory agencies work together to create a framework for regulating AI.
Here are some of the key takeaways:
- Risks associated with AI are part of its evolution.
"For instance, some writers fear that autonomous weapons could end up being genocidal war machines. Other times, AI has been shown to exhibit racial or gender bias and makes faulty financial decisions.”
- Fears of AI making mistakes is overblown. Humans are error prone too.
"The key difference is that when an expert makes a mistake, the human expert is responsible for it. As of now, the same liability does not apply to AI."
- AI is moving at a faster pace than regulatory agencies can keep up.
“Due to the novelty of the technology and the pace of change, however, progress has felt slow and ad-hoc, with many regulatory bodies apparently taking a wait-and-see approach.”
- Regulating AI starts with transparency, auditability, and accountability.
“Business leaders should at a minimum be able to show how AI is used by experts to make critical decisions (transparency), while also demonstrating that they’re able to work back from the final decision through the processes, technologies, and data that were used to arrive at those decisions (auditability).”
Read the full post on the Forbes Technology Council.
- CCPA: Will Your Business Be Responsible For Your Data Service Provider? https://compliance.ai/blog/ccpa-will-your-business-be-responsible-for-your-data-service-provider/
- Why AI Needs Experts https://www.compliance.ai/blog/why-ai-needs-experts/
- Compliance.ai's Top 6 Industry Predictions for 2019 https://www.compliance.ai/blog/complianceai-top-6-industry-predictions-for-2019/