Anthropic, a rising player in the field of artificial intelligence (AI), has made its mark by launching its new AI chatbot named Claude 2 to the public. With a focus on responsible AI development, the startup has attracted significant attention from major investors and companies such as Google, Salesforce, and Zoom.
Anthropic, founded in 2021 by former executives of OpenAI, has been steadily investing in its AI capabilities to compete with tech giants like Microsoft-backed OpenAI and Google. The race to develop advanced chatbots powered by generative AI has intensified, prompting Anthropic to introduce Claude 2 after raising an impressive $750 million in two financing rounds.
The company has been testing its AI models with prominent businesses like Slack, Notion, and Quora for the past two months. This period of testing has allowed Anthropic to refine and improve the safety features of Claude 2, resulting in what they believe to be the safest version of the chatbot so far. Anthropic has also accumulated a significant waitlist of more than 350,000 people eager to access Claude 2’s application programming interface and consumer offering.
Initially, Claude 2 will only be available to users in the United States and the United Kingdom, but the company plans to expand its availability to other regions in the future.
Generative AI chatbots, such as OpenAI’s ChatGPT, have gained significant attention in the tech world due to their ability to provide sophisticated and conversational responses to text prompts. Despite concerns about bias propagation associated with this technology, chatbots have found applications in various sectors ranging from education and healthcare to online advertising and travel.
OpenAI’s release of GPT-4 earlier this year marked a major update to the underlying technology behind ChatGPT. The inclusion of internet browsing capabilities in ChatGPT was a significant advancement, but concerns about unintended uses led to the feature being disabled.
Likewise, Google introduced its competitor, Bard, in February. Since then, the tech giant has further improved the chatbot’s skills in mathematics and coding, integrating it into more of their products.
Anthropic’s ambitions match those of its competitors. With Claude 2, the company claims to have developed a chatbot capable of summarizing up to approximately 75,000 words, equivalent to the length of a book. This surpasses the capabilities of ChatGPT, which can handle only around 3,000 words.
Daniela Amodei, co-founder of Anthropic, stated that developing Claude 2 involved at least two months of intensive work, with a specific team of 30 to 35 people solely dedicated to the AI model, supported by a total of 150 individuals. Amodei emphasized the rapid growth of the AI market and the ample opportunities for various players to succeed within it.
Anthropic’s achievements in the AI industry have also earned them recognition from policymakers. In May, the company was invited along with Google parent Alphabet, Microsoft, and OpenAI to a meeting at the White House to discuss responsible AI development with Vice President Kamala Harris.
The company’s impressive progress was reflected in its recent funding rounds. In March, Anthropic raised $300 million, followed by an additional $450 million in May, valuing the startup at $4.1 billion.
Amodei shared concrete examples of Claude 2’s improvements compared to its previous version. The upgraded chatbot scored higher on the Bar exam’s multiple-choice section, from 73% to 76.5%, and demonstrated improved performance in a Python coding test, with a score of 71% compared to its prior iteration’s score of 56%.
While Claude 2 has shown remarkable progress, Amodei acknowledged that challenges persist in the field of AI chatbots. One common issue is the tendency of chatbots to produce incorrect answers, known as hallucinations among tech companies. Amodei emphasized that no language model is completely immune to such errors, and Claude 2 is no exception.
Anthropic’s foray into the AI chatbot market highlights the immense demand for advanced language models. As the industry continues to develop, it is crucial to address ethical considerations while ensuring the safety and reliability of these AI applications.
