AI’s Trust Problem

Twelve persistent risks of AI that are driving skepticism.

Bhaskar Chakravorti. Harvard Business Review. May 03, 2024. Part 1 of 3.

Summary.

As AI becomes more powerful, it faces a major trust problem. Consider 12 leading concerns: disinformation, safety and security, the black box problem, ethical concerns, bias, instability, hallucinations in large language models, unknown unknowns, potential job losses and social inequalities, environmental impact, industry concentration, and state overreach. Each of these issues is complex — and not easy to solve. But there is one consistent approach to addressing the trust gap: training, empowering, and including humans to manage AI tools.

AI’s Trust Problem refers to the challenges and concerns surrounding the trustworthiness of artificial intelligence (AI) systems. It encompasses issues related to reliability, transparency, accountability, ethical alignment, and user confidence in AI technologies. As AI becomes increasingly integrated into various aspects of society—from healthcare and finance to transportation and personal assistants—the importance of establishing and maintaining trust in these systems grows exponentially.

AI’s Trust Problem is a multifaceted challenge that encompasses technical, ethical, societal, and regulatory dimensions. Addressing this problem is crucial for the widespread adoption and positive impact of AI technologies. By enhancing transparency, ensuring accountability, promoting fairness, strengthening security, building user confidence, and establishing ethical guidelines, stakeholders can work towards overcoming the trust barriers associated with AI. This collective effort is essential to harness the full potential of AI while mitigating its risks and ensuring it serves the greater good.

With tens of billions invested in AI last year and leading players such as OpenAI looking for trillions more, the tech industry is racing to add to the pileup of generative AI models. The goal is to steadily demonstrate better performance and, in doing so, close the gap between what humans can do and what can be accomplished with AI.

There is another gulf, however, that ought to be given equal, if not higher, priority when thinking about these new tools and systems: the AI trust gap. This gap is closed when a person is willing to entrust a machine to do a job that otherwise would have been entrusted to qualified humans. It is essential to invest in analyzing this second, under-appreciated gap — and in what can be done about it — if AI is to be adopted widely.

The AI trust gap can be understood as the sum of the persistent risks (both real and perceived) associated with AI; depending on the application, some risks are more critical. These cover both predictive machine learning and generative AI. According to the Federal Trade Commission, consumers are voicing concerns about AI, while businesses are worried about several near to long term issues. Consider 12 AI risks that are among the most commonly cited across both groups:

• Disinformation

• Safety and security

• The black box problem

• Ethical concerns

• Bias

• Instability

• Hallucinations in large language models

• Unknown unknowns

• Job loss and social inequalities

• Environmental impact

• Industry concentration

• State overreach

Taken together, the cumulative effect of these risks contribute to broad public skepticism and business concerns about AI deployment. This, in turn, deters adoption. For instance, radiologists hesitate to embrace AI when the black box nature of the technology prevents a clear understanding of how the algorithm makes decisions on medical image segmentation, survival analysis, and prognosis. Ensuring a level of transparency on the algorithmic decision-making process is critical for radiologists to feel they are meeting their professional obligations responsibly — but that necessary transparency is still a long way off. And the black box problem is just one of many

Episode 3133 - October 26 - Tiếng Anh - Phần 1 của 3 - Vấn đề Niềm tin đối với AI - Vina Technology at AI time

Vina Technology at AI time - Công nghệ Việt Nam thời AI

Episode 3133 - October 26 - Tiếng Anh - Phần 1 của 3 - Vấn đề Niềm tin đối với AI - Vina Technology at AI time