ChatGPT | Microsoft advances 10 billion US, fraudsters flock

Already seen as a serious competitor to Google’s search engine, ChatGPT could benefit from an additional investment of 10 billion US from Bing’s owner, Microsoft. Perverse effect of the popularity of this conversational robot: dozens of fraudulent applications have appeared in recent days.

According to information reported Monday evening by the American media Semafor, citing “sources familiar with the matter”, Microsoft is preparing to invest 10 billion US in the San Francisco company behind ChatGPT, OpenAI, which would give it a valuation of 29 billion US (about 39 billion CAN). Under a complex arrangement, Microsoft would take 75% of the profits until its investment was recouped, after which it would get 49% of OpenAI’s shares.

Neither Microsoft nor OpenAI reacted to this news on Tuesday.

Microsoft’s move is hardly a surprise in the tech industry, as the company’s interest in this amazing chatbot has been reported by multiple outlets for months. Microsoft had already invested US$1 billion in July 2019 in OpenAI, whose chatbot is hosted on the Microsoft Azure cloud platform.

For Microsoft, it is above all a question of positioning its own search engine, Bing, as the replacement for Google. Despite numerous investments and aggressive promotion by Microsoft for Bing, this search engine capped in terms of market share in second place with 8.9%, far from the 88% held by Google, according to Statista.

Plagiarism and virus

ChatGPT, a chatbot based on a third version, is a text generator born in 2020 trained on billions of texts reflecting human opinions and creations until 2021. It is not connected to the Internet, which makes it prevent us from answering burning topical questions. Since its release to the general public on November 30, it has aroused fascination and concern about its ability to imitate human written responses, in some forty languages, including French.

This robot can answer philosophical, political or cultural questions with great plausibility as well as create works on requested themes. Its quality is credible enough that fears related to its use for plagiarism have been raised.

More troubling, researchers from a cybersecurity lab, Check Point, revealed earlier this week that ChatGPT was already being tested by hackers for its ability to write computer code, fix script errors, and help build software. malicious software.

Victim of his own success

The chatbot has also been used as the bait for dozens of fraudulent apps, offered on both the Apple App Store and the Google Play Store. Using the robot name in the middle of a longer name like ChatGPT AI with GPT3 Goat Chatthese applications have climbed into the top 10 most popular downloads in Canada. They imitate the operation of the real ChatGPT in an astonishing way, even in French, according to the tests attempted by The Press. It is not known whether these apps draw on ChatGPT’s own programming or whether they are convincing imitations.

The big difference is that after a certain number of interactions, the apps offer paid plans, like ChatGPT Pro, offered for US$9.99 per week.

However, the real ChatGPT is free and is only available on a web browser, from the address openai.com/blogue/chatgpt/. OpenAI has not yet released a mobile application. The conversational robot is however a victim of its success, when it was almost impossible to connect to it on Tuesday due to too much demand.

As of this writing, some scam apps including ChatGPT Chat GPT AI With GPT-3, were no longer offered on the App Store. Neither Apple nor Google had officially reacted to the presence in their virtual store of these fraudsters.