A new report from Surfshark suggests that the number of AI incidents in 2023 is set to surpass all previous records.
Based on the AI Incident Database, the study found that AI incidents have exponentially risen since 2010, identifying that 2023 already has more than half the number of AI incidents reported in all of 2022, and it is set to smash all previous records.
Commenting on the potential dangers of this steep increase, Surfshark’s Elena Babarskaite said: “Artificial Intelligence incidents are not only growing year-by-year but also drawing more public attention.
“A recent prominent example is the viral deep fake image of Pope Francis in a white puffer jacket, demonstrating the astonishing realism achievable with AI image generators.
“However, the implications go beyond amusement, as AI-enhanced fake news could continue to mislead the public in the future. Other, more severe AI incidents can go as far as enabling racism, violence, or crimes.”
What is an AI incident?
The AI Incident Database defines an AI incident as an alleged harm or near harm event to people, property, or the environment in which an AI system is implicated.
Whereas an AI issue is an alleged harm or near harm by an AI system that has yet to occur or be detected.
The database, which collects, verifies, and classifies user-reported events in which harm was either caused or almost caused by AI, logged its first incident in 1983.
Here, one of the Soviet Union’s automated systems incorrectly identified incoming missiles, which almost ended in a nuclear catastrophe.
Since that doomsday scenario, AI technology has advanced dramatically and is now used in various aspects of our daily lives, such as smart appliances, navigation apps, facial recognition, and security.
However, this significant rise in AI adoption has also welcomed an increase in annual incidents.
Mapping the rise of incidents
For the study, data was collected between 2010 and 16 June 2023. The study characterises 2010 to 2014 as the initial stage of AI in which the technology started to become relevant, seeing just 18 incidents, with an average of five per year. This accounts for just 3% of all incidents.
The growth of AI stage stretched from 2014 to 2019 and saw the development of software advancements that enabled companies to make major investments in AI research, helping to advance sectors such as computing and biotech.
This stage saw 207 incidents, averaging 35 annually, which accounted for 40% of all reported cases.
The most recent stage, characterised by the adoption of AI, saw the development of GPT-3 by OpenAI in 2020, the predecessor of ChatGPT.
This stage has experienced 293 incidents, with an average of 79 per year and making up 57% of all incidents.
In 2022 alone, there were 91 incidents, with 55 in 2023 so far. This means that at this current rate, the number of AI incidents in 2023 will surpass all previous records.
