收听本期播客
阅读正文
A significant rivalry is unfolding in the artificial intelligence sector, with tech giants Anthropic and OpenAI engaged in a public debate, often dubbed an ‘ad war.’ This dispute centres on the future of advertising within their popular AI products and highlights fundamental differences in their business models and ethical approaches to AI development.
Recently, Anthropic launched an advertising campaign for its Claude AI chatbot, which playfully criticises the concept of other AI platforms embedding targeted advertisements directly into user conversations. For instance, one advertisement amusingly shows a chatbot offering intrusive and irrelevant advice, such as recommending shoe insoles to someone wishing to appear taller, or suggesting a dating site to improve communication with their mother. Anthropic’s message is clear: ‘Ads are coming to AI. But not to Claude,’ a statement implicitly aimed at OpenAI, creators of ChatGPT, who have announced plans to integrate advertisements.
Sam Altman, CEO of OpenAI, quickly responded, describing Anthropic’s campaign as ‘clearly dishonest.’ He strongly denied that OpenAI would ever implement ads in such an intrusive manner. Altman stated that OpenAI’s strategy is to make AI accessible to billions globally who cannot afford subscription services. He assured the public that their advertisements would be ‘separate and clearly labeled,’ guaranteeing they would not influence the chatbot’s responses. Furthermore, OpenAI has committed not to share user conversations with advertisers and will provide options for users to disable personalisation or upgrade to an ad-free paid subscription.
This shift in OpenAI’s approach is noteworthy, as Altman had previously considered advertising a ‘last resort’ in late 2024. However, facing increased investments in AI infrastructure and a slowdown in new subscriber growth, the company appears to be seeking alternative revenue streams. Anthropic, founded by former OpenAI researchers with concerns about AI safety, argues that incorporating ads would prevent Claude from serving as a truly helpful assistant for professional tasks and deep thinking, especially considering the often personal nature of AI interactions.
The ongoing ‘ad war’ raises crucial questions regarding user privacy and the potential misuse of user vulnerabilities, particularly when individuals might consult AI on sensitive subjects like health. While targeted advertising could potentially help regulate AI content by encouraging corporations to withdraw support from harmful material, the overall impact on user trust remains uncertain. Only time will tell if OpenAI’s decision will lead users to ad-free competitors such as Claude.