ChatGPT gets competition! Say hello to ‘Claude 2’, AI chatbot developed by Anthropic PBC

Anthropic PBC, an American AI-based company and public-benefit corporation created by former OpenAI members, has introduced Claude 2, a new ChatGPT competitor. It has launched Claude 2 for public access in the United States and the United Kingdom, at a time when debates about the safety and societal impact of artificial intelligence are becoming more intense.

This latest version follows the release of its predecessor, Claude, which was exclusively available to companies just five months ago. Unlike its earlier iteration, Claude 2 is accessible through a public beta website and an API.

Anthropic vs. ChatGPT

According to Ethan Mollick, an Associate Professor at the Wharton School of the University and one of the chatbot's beta testers, Claude 2 offers two significant advantages over other models. Firstly, it excels at handling documents, particularly PDFs, which GPT struggles with, displaying a sophisticated understanding of them.

Additionally, it maintains the most pleasant AI personality. However, Mollick advises caution when using the model for data-related tasks, despite its acceptance of CSV (comma-separated values) files. Unlike the Code Interpreter, Claude 2 tends to generate inaccurate responses.

In May, Anthropic extended the length of input and output for Claude. This enhancement enables the chatbot to process lengthy technical documents or even entire novels. Claude is also capable of generating longer documents such as memos, letters, and stories.

Furthermore, the chatbot can get up-to-date information from the web, a feature that is limited to paying users of ChatGPT.

Is Claude 2 scoring more than ChatGPT?

Anthropic's latest AI assistant, Claude 2, achieved impressive scores on various assessment exams. According to the announcement blog, it scored 76.5 percent on the multiple-choice section of the Bar exam and performed in the 90 percentile on the reading and writing portions of the GRE.

Similar to its popular competitor, Claude offers additional benefits. In addition to improved text-based responses, Anthropic has enhanced Claude's coding abilities. The company reports that the bot achieved a higher score of 71.2 percent on a widely used Python coding test.

Furthermore, Claude achieved an 88 percent on a comprehensive set of primary school math problems in the United States.

Claude 2 approved by the UN

The remarkable success of the chatbot has raised concerns among experts and lawmakers regarding its potential risks to privacy and society. Italy, for instance, banned ChatGPT in March due to concerns about data safety, and prominent figures in the tech industry have called for a pause on advanced AI systems.

To address these concerns, Anthropic emphasises that Claude is built on trust and safety practises derived from major organisations. The chatbot was trained using various documents, including the UN Declaration of Human Rights and Apple's terms of service. It has been actively working to improve Claude 2's underlying safety, making it less likely to generate offensive output when prompted.

Additionally, Dario Amodei, the CEO of Anthropic, met with Britain's Prime Minister Rishi Sunak in May as part of a delegation of AI leaders. The discussions revolved around the risks posed by AI, ranging from disinformation and national security to threats.