**AI Incidents on the Rise: A Growing Concern for Society**
AI incidents have been on the rise since 2010, with a projected increase in 2023 surpassing the number of incidents in 2022. While some incidents may seem minor, others have had severe consequences, such as the incident in 1983 that nearly led to global nuclear war. In recent years, we have witnessed AI-enabled malfunctions, errors, fraud, and scams that have impacted various aspects of our lives. From deepfakes used to influence politics to chatbots providing harmful health information and self-driving vehicles endangering pedestrians, the impact of AI incidents cannot be overstated.
**Surfshark’s Report: Identifying the Culprits**
According to security company Surfshark, Tesla, Facebook, and OpenAI have been identified as the worst offenders, collectively responsible for 24.5% of known AI incidents. These incidents range from Facebook’s algorithms failing to detect violent content to scammers using deepfake technology to defraud users. Tesla’s AI issues primarily stem from its Autopilot and Full Self-Driving software products, which have been involved in accidents due to unexpected braking or failure to detect objects in front of them. OpenAI faces challenges related to data privacy and has even allegedly made death threats against users in Bing Search.
**The Alarming Growth of AI Incidents**
The rise in AI incidents can be attributed to the exponential increase in AI investment and usage. Chatbots, in particular, have experienced a staggering 261% increase in search traffic from February 2022 to February 2023, according to software review service G2. Additionally, AI products are among the fastest-growing software products in G2’s database, indicating the growing reliance on AI technology. While skepticism surrounds the use of AI systems among the general population, a recent survey of 1,700 software buyers conducted by G2 revealed that 78% of respondents trust the accuracy and reliability of AI-powered solutions.
**The Soaring Numbers: A Cause for Concern**
Despite the trust placed in AI systems, the number of AI incidents has witnessed a significant jump in recent years. From an average of 10 incidents per year between 2014 and 2022, the past few years have witnessed an average of 79 major AI incidents annually, representing a growth rate of 690% in just six years. This growth shows no signs of slowing down, as 2023 has already seen half the number of incidents reported in the entirety of 2022.
**The Dark Side of AI: Implications and Challenges**
The consequences of AI incidents go beyond amusement, as demonstrated by the viral deepfake image of Pope Francis wearing a white puffer jacket. The realistic nature of such AI-generated content raises concerns about the potential for AI-enhanced fake news to mislead the public and facilitate racism, violence, or even crimes. Notably, a significant number of AI incidents involve the word “black,” such as the wrongful arrest of a Jefferson Parish resident due to facial recognition technology falsely identifying him as another individual. Moreover, Facebook faced backlash after its AI tools erroneously labeled black men as “primates.”
Safety and equity are not always prioritized in the rush to incorporate AI into various domains. Bias in AI systems is gradually being addressed, but it remains a persistent issue. As the reliance on AI continues to grow, it is crucial to ensure that proper safeguards are in place to mitigate AI incidents and promote fairness and security.
The increasing prevalence of AI incidents is a cause for concern, given their potential to disrupt global affairs, deceive the public, and endanger lives. Companies like Tesla, Facebook, and OpenAI must address the challenges they face and take proactive measures to prevent future incidents. At the same time, society as a whole must prioritize safety and equity when implementing AI technology. With the right precautions and responsible usage, AI can continue to benefit humanity without compromising our well-being.