Artificial intelligence startup Anthropic has unveiled a new version of its AI chatbot, Claude, in a bid to position itself as a provider of safer conversational AI. The latest release, called Claude 2, is now available for use by individuals in the US and UK via claude.ai, while businesses can access it via an application programming interface. This launch follows the earlier introduction of an initial version of Claude to select companies, including Quora, which integrated the chatbot into its question-answering app called Poe.
Similar to its predecessor, Claude 2 is built on a robust language model and is capable of performing various written tasks, such as summarizing, searching, answering questions, and coding. Both versions have the ability to process large amounts of text, enabling users to ask the bot to summarize entire books, for example. However, Claude 2 boasts an advantage over its predecessor in terms of generating longer responses, with the potential to reach up to approximately 3,000 words per response. According to the company’s data, Claude 2 is also expected to provide more accurate answers in certain domains like coding and grade-school-level mathematics.
Anthropic’s primary objective with Claude has been to create a chatbot that is less susceptible to manipulation compared to its counterparts. The company seeks to control the generation of harmful or offensive output in response to prompts, addressing an issue that has plagued other chatbots, including OpenAI’s ChatGPT. Anthropic claims that its tests indicate Claude 2 outperforms the existing model in generating harmless responses to harmful prompts.
Daniela Amodei, Anthropic’s President and Co-founder, acknowledged that while Claude 2 is an improvement over the older version, it still has limitations and falls short of perfection. The San Francisco-based startup, established in 2021 by a group of former OpenAI employees, including Daniela and her brother Dario Amodei, has been vocal about the need for caution and responsible development in the field of AI.
In terms of funding, Anthropic recently raised $450 million, bringing its total funds raised to over $1 billion. The company plans to expand the availability of Claude to more countries in the coming months, as it continues to emphasize its commitment to building secure and reliable AI systems.
