Potential Multi-Million Dollar AI Training Deals Being Explored by Apple with News Publishers
Apple is in discussions with prominent publishers, such as NBC News and Condé Nast, to acquire news archives to improve its artificial intelligence systems.
According to sources who requested anonymity to discuss delicate discussions, the technology giant has offered multiyear deals worth at least $50 million to license news article archives. Apple approached Condé Nast, publisher of Vogue and The New Yorker, NBC News, and IAC, owner of People, The Daily Beast, and Better Homes and Gardens.
Neural networks are constructed by recognizing patterns in large photo or text datasets. By examining thousands of cat photographs, a computer can recognize cats. Microsoft, OpenAI, Google, Meta, and others have created chatbots and other products using the technology. The tools have the potential to transform work and create billions in sales.
Three people involved with the negotiations said Apple’s demands were too expansive for certain publishing executives. The initial presentation involved wide licensing of publishers’ archives, with publishers possibly liable for Apple’s legal responsibilities.
Also Read – Evaluation Of Sonodyne Malhar: An Indian Masterwork Catering To Purists In The Art Of Pure Sound
The insiders also said Apple was unclear about how it will use generative A.I. in the news industry, which might be a competitive issue given its large device viewership. Despite its rapid growth, generative AI raises worries about training biases, especially in sensitive fields like politics.
Understand these crucial points:
Data prejudices:
AI model data often reflects social prejudices. This means a model trained on news stories may accidentally hold political biases and favor specific beliefs. Even with balanced data, training algorithms might generate unintentional biases. This might happen when the system prioritizes or weights information, boosting some opinions and suppressing others.
Output bias:
AI outputs can be biased. In a political speech, the AI may favor specific candidates or policy stances depending on its training biases.
Impact on Society:
AI political prejudice can hurt society. It can cause unfair discrimination, public opinion manipulation, and silencing of diverse opinions. Biased AI in social media algorithms may stifle political viewpoints or spread misinformation.
Addressing Concerns:
Researchers and developers are actively reducing these biases. Different methods include:
- Diverse and representative training data exposes the AI to a variety of perspectives and information.
- Creating debiasing algorithms to discover and rectify training data and algorithm biases.
Transparency and explainability:
Making the AI’s decision-making process more public can assist in discovering and removing biases.
Remember that AI is new and growing. While bias issues are genuine, research and development are underway to solve them. Responsible and impartial methods are essential as we build and perfect AI.
