The trend: Progress in AI manufacture the technology omnipresent, particularly in the financial sector, but questions of ethics and reliability weigh heavily. Surveys reveal that C-level executives in several industries aren’t too eager to start using it themselves.
- AI is commonly used for lower level tactical decision making, but it is not a common practice for higher level strategic decisions where leaders prefer to rely on their instincts, by Harvard business review.
- A survey of 2,190 executives from nine countries, representing banking/finance, insurance, telecommunications, retail, HC/LC and governmentfound that 67% of CEOs often prefer to make decisions based on their intuition and experience rather than information from data analysis, by KPMG.
- In a Deloitte survey, 67% of executives said they were not comfortable accessing or using data from advanced analytics systems.
- 42% data scientists of a representative sample of industries said their results are not used by business decision makers.
AI is easy to implement, but guarantees are not: Since assess loan applications and determine insurance premiums to deploy customer service chatbotsAI affects our world, with much of it happening in the background.
However, recommendations for overcoming issues that make leaders wary, such as creating more reliable AI models, avoiding data biasensure ethics and explainability – are not easy to achieve.
- The increasing complexity of AI, including Deep Generative Models (DGMs)) created by merging neural networks and generative modeling, makes it even more opaque, limiting developers’ ability to reduce bias and ensure consistency of actual results.
- Efforts to teach ethics to AI, effectively endowing it with a superego, result in such a dilemma that it might be impossible. AI expert Lance Eliot declared“The more AI gets better, the worse the AI ethics problem gets.”
- AI prowess is downright sinister sometimes, and ironically, despite its promise of improved cybersecurity, AI makes banks vulnerable to Russian hackers.
The bigger picture: Widespread concern about AI is at odds with the cavalier approach to its implementation and makes last year dismissal AI ethicists shake. AI innovations suggest the technology could be applied in potentially endless ways, but that’s not necessarily the best use case.
- Implementing AI already affects the lives of millions of people, and using it for high-level strategic decisions carries outsized risk.
- Technology does not need to be used in every way imaginable for society to reap the benefits.
- AI in collaboration with robotics and automation could potentially work in all professions, but the reality is that many people value human contributions only.