Anthropic, a competitor of ChatGPT, launches enhanced models to empower its Claude AI chatbot

Anthropic, a leading company in generative artificial intelligence, revealed on Monday that it has introduced new models to enhance its Claude AI chatbot, amidst increasing competition for ChatGPT.

The company said the three new AI models – called Claude 3 Opus, Sonnet and Haiku – were its most powerful tools to date and led the industry in their ability to match human intelligence.

Founded in 2021, Anthropic was created by former employees of OpenAI, the company behind ChatGPT, and has been funded by Google and partnered with Amazon to develop new technology.

The company has made its mark by releasing AI models that aim to set tighter guardrails than those behind ChatGPT and other chatbot competitors.

But that approach has faced backlash since Google’s Gemini model was released last month and was criticized for confusions such as creating images of the ethnically diverse ranks of World War II Nazi troops.

Some industry observers also complain that chatbots have become less effective as companies introduce tighter controls in response to controversy over the technology running away or giving wrong answers.

Anthropic acknowledged that safeguards can go too far, and said the new designs would avoid “unnecessary opt-outs” that were a problem in its previous releases.

“The Opus, Sonnet and Haiku refuse to respond to prompts confined to the system’s guardrails, significantly less than previous models,” it said.

Anthropic said its model, the Opus, was the most powerful of the three and was able to beat the others in key benchmarks, including math.

Considered one of the largest AI chatbot manufacturers, Claude is closely allied with Amazon and its AWS cloud division, which provides the company’s compute-intensive needs.

It has also received investments from Google and other Silicon Valley heavyweights.

Unlike its competitors, Anthropic’s Claude chatbot does not generate images and only allows users to use images as requests for analysis.

Competing tools from OpenAI and Google generate images on demand, but Anthropic executives believe customers won’t demand the feature.

Like other AI giants, Anthropic faces lawsuits from content creators who accuse the company of stealing copyrighted material to build its models.

Universal and other music publishers sued Anthropic in a US court last year for using copyrighted lyrics to train its systems and generate answers to users’ questions.