Why AI Needs Experts

Why AI Needs Experts
Why AI Needs Experts

Compliance.ai’s CEO and Co-Founder, Kayvan Alikhani,  discusses the new opportunities that AI will bring and explains why AI requires human expertise, now more than ever in his article, published on Forbes Technology Council.

Here are some of the key takeaways:

  • Fear of AI is on the rise.
    • According to a Gallup poll from last year, a majority of Americans (58%) believe AI is a bigger threat to their jobs than immigration and offshoring. But have no fear: many of the jobs that people fear will be eliminated, such as assembly line positions, have already been outsourced before AI took off.
  • AI and ML will impact knowledge workers
    • According to the Brookings Institute, Artificial Intelligence (AI) and machine learning (ML) will impact knowledge workers, but not necessarily in the way you may expect. In fact, the World Economic Forum predicts that AI and robotics will create 60 million more jobs than they will eliminate by 2022.
  • Supervised Learning will be required
    • Specifically with ML, unsupervised learning requires the help of humans to fine-tune their algorithms and assess their confidence levels. Without human intervention, unsupervised learning algorithms are only about 80% accurate, far below the standard required for many businesses. This model of learning is called human-in-the-loop (HITL).
  • Expert-in-the-loop will be the gold standard approach
    • HITL is best for getting a high volume of data, but when it comes to domain expertise businesses are turning to expert-in-the-loop (EITL) where domain experts are recruited to train AI in specialized, heavily regulated fields.
  • AI will free professionals from mundane tasks
    • One study found that law firms lose more than $9,000 per year per lawyer on document management activities. This translates into a 10% loss in productivity that could have gone to value-creation activities instead.

Read the full post on the Forbes Technology Council.

Other Resources:

NEW: How RegTech is Tackling Today’s Complex Regulatory Landscape Better Than Legacy SystemsFind Out

X