Security Company: Tesla, Facebook and OpenAI are responsible for 24.5% of ‘AI incidents’
The first \”AI event\” nearly caused nuclear war. Recent AI-enabled scams, frauds, and errors include deepfakes that were used to influence political decisions, bots giving bad health advice, and self-driving cars that endanger pedestrians.
According to the security company Surfshark’s report, Tesla, Facebook and OpenAI are responsible for 24.5% out of all AI incidents that have been reported.
In 1983, a Soviet automated system believed it had detected nuclear missiles coming from the United States. This almost led to a global conflict. Surfshark reports that this is the first incident (though there’s debate over whether an automated system dating back to the 1980s can be considered artificial intelligence). The National Eating Disorders Association’s (NEDA) chatbot Tessa was shut down after it gave people who were seeking help with eating disorders dangerous advice. A self-driving Tesla failed to notice a pedestrian, then broke the law when it did not yield to a person at a crosswalk. Another incident involved a Jefferson Parish resident who was wrongfully arrested in Louisiana after a facial recognition software developed by Clearview AI mistook him as another individual.
Source:
https://www.forbes.com/sites/johnkoetsier/2023/06/28/security-company-ai-incidents-up-690-worst-offenders-are-tesla-facebook-openai/?sh=2cfc485b5b12
