Apple has reportedly disabled the functionality of summarizing news articles using artificial intelligence on the beta version of a new iOS operating system due to inaccuracies and misinformation. Previous attempts to rectify the issue were unsuccessful, leading Apple to take action. The decision came after complaints from news organizations whose articles were being misrepresented or falsely attributed by the AI-powered feature. The latest iOS update, version 18.3, is set to be released soon, with improvements expected for the news feature.
Concerns over the feature’s credibility arose in November when headlines and summaries were found to be misleading and inaccurate. Instances included false reports of individuals committing crimes and winning competitions in news alerts. Vincent Berthier, a technology and journalism expert at Reporters Without Borders, urged Apple to remove the feature to protect the credibility of media outlets and ensure the public’s access to reliable information.
The issue of AI generating misleading information, known as “hallucinations,” has become a significant challenge for companies relying on AI for decision-making. Examples of fabricated content produced by AI systems highlight the risks associated with automated technologies. OpenAI’s Whisper transcription software was recently criticized for inserting fake text into conversations, including potentially harmful content.
Amazon also faced challenges with AI hallucinations during its efforts to enhance Alexa into a smarter voice assistant. The incorporation of generative AI technology into Alexa’s core functionality has presented hurdles, including the risk of misinformation being generated. Addressing these issues is vital for ensuring the accuracy and reliability of AI-driven technologies in various industries.
For comprehensive coverage of AI-related topics, readers can subscribe to the daily AI Newsletter provided by PYMNTS. This curated newsletter offers insights and updates on the latest developments in artificial intelligence and its applications in various sectors.
Vocabulary List:
- Functionality /ˌfʌŋkʃəˈnælɪti/ (noun): The specific function or purpose of a system or feature.
- Inaccuracies /ɪˈnækjərəsi/ (noun): Errors or mistakes that lead to something being incorrect or untrue.
- Misinformation /ˌmɪsɪnˈfɔːr.mə.ʃən/ (noun): False or misleading information.
- Credibility /ˌkrɛdəˈbɪlɪti/ (noun): The quality of being trusted and believed in.
- Hallucinations /həˌluːsɪˈneɪʃənz/ (noun): Instances of perceiving things that are not present often used in the context of erroneous outputs from AI.
- Fabricated /ˈfæbrɪkeɪtɪd/ (adjective): Constructed or manufactured often in a deceptive manner.
How much do you know?
