ChatGPT has taken the tech world by storm in recent weeks, but it’s not the only one in the fray. Anthropic, an artificial intelligence company founded by former OpenAI employees announced Claude, an AI similar to ChatGPT looking to gain a foothold in the market. Although the company may sound unfamiliar to many, Anthropic has staff with years of experience and its technology is present in Notion, Quora and DuckDuckGo.
In accordance with TechCrunch, Claude is very similar to ChatGPT in that it can write text, source code, analyze and summarize documents. What differentiates this chatbot from its competitor is that makes fewer mistakes and is not as toxic. Anthropic assures that it is easy to talk to him, since he does not lose his head so easily.
Claude trained in the same way as ChatGPT, so uses a database of web pages published up to 2021. Anthropic said they have taken special care to avoid racist and sexist responses. The latter is important, as ChatGPT has shown a biased and sometimes offensive trend in queries.
The idea of a more manageable AI is based on the concept of Constitutional AI, a technique that Aligns AI systems with human intentions. According to Anthropic, Claude answers the questions using a set of 10 guiding principles. The company says these principles respect freedom of choice, maximize positive impact, and refrain from offering harmful advice.
Claude wants to gain ground in a segment dominated by ChatGPT
While ChatGPT dominates the scene thanks to its integration with Bing, Snapchat, Discord, and coming soon to GM cars, Claude is gaining ground. Anthropic confirmed that its artificial intelligence is behind Notion AIa new feature that automates tasks and writing notes in the popular productivity suite.
Claude is also part of DuckAssist, the DuckDuckGo assistant that combines the AI of OpenAI and Anthropic in one product. The search engine has endowed its Instant Answers function with artificial intelligence. This means that Claude and ChatGPT will offer automatic suggestions when you ask a question.
Lastly, Anthropic confirmed that Claude is behind Poe, the Quora chatbot that answers user questions and that is available within the mobile application. The rest of the list includes clients like RobinAI and AssemblyAI, who take advantage of the chatbot’s capabilities to translate and explain complex concepts in simple ways.
Anthropic’s AI Isn’t Perfect (Yet)
claude It has been in a testing phase for a few months and is ready to go on the market. Anthropic confirmed that they are confident in meeting the demand and invited organizations to request access to their AI. The chatbot will be offered via an API in two variants: Claude and Claude Instant, although the cost of each has not been confirmed.
Although the competition from ChatGPT is gaining ground, OpenAI doesn’t have to worry. On paper, Claude was designed under a series of principles that will avoid toxic or biased responses, but in reality It carries similar problems as its counterpart. Anthropic’s chatbot is not very good at math and also makes up answers or suggests illegal activities.
The good news is that its creators will continue to adjust the model to offer a reliable alternative in the future.