Half of the copies of students from a Lyon faculty written by artificial intelligence, Microsoft announcing that it wants to invest 9 billion euros, an application capable of detecting the first signs of Alzheimer’s disease… ChatGPT, the application based on “generative” artificial intelligence (AI) foreshadow a revolution?
Since the beginning of the year, the digital world has been talking only about ChatGPT. Developed by the Californian start-up OpenAI, this softwareAI panics the counters and leads to many effects of announcements and disruptive applications.
Microsoft announces that it wishes to invest up to 9.34 billion euros in this solution to improve Office in particular. After forging a first partnership with this small company in 2019, its investment would value OpenAI at $29 billion.
For the IT giant, ChatGPT could boost the capabilities and performance of its office suite by being able to generate documents without human intervention. For the start-up, the cloud and Microsoft datacenters would allow it to optimize the use of its software, which is very greedy in computing resources.
All sectors concerned
But why such excitement around ChatGPT? This software demonstrates how AI would become essential by moving from “analytical” AI to “generative” AI.
In the first case, computers can analyze a set of data and find templates for a multitude of use cases, whether it’s detecting fraud or spam, predicting when your delivery will arrive, or predicting the next video TikTok to show you.
In the second case, “generative” AI, computers generate something new rather than analyzing something that already exists. This idea is not new, but “generative” AI is set to become not only faster and cheaper, but also better in some cases than human beings.
All sectors requiring human beings to create original works (video games, advertising, architecture, computer coding, etc.) could be strongly impacted by this major development. Some functions could be completely replaced by generative AI, while others are more likely to thrive on the human/machine pairing.
Hackers take advantage of ChatGPT
Different areas are therefore concerned. Text is the most advanced, but natural language is still difficult to master. Today, templates are good enough for writing uncomplicated texts. The creation of images is more recent, but it is beginning to be used more and more on social networks.
Speech synthesis has existed for several years, especially on smartphones and connected speakers. These are still consumer applications. Voice generation for movies and podcasts is not yet mastered to obtain human-quality speech that is not mechanical. Finally, work is carried out in biology and chemistry.
This first wave of generative AI apps resembles the one we experienced when the first apps appeared on the first smartphones. There are still many problems to be solved concerning in particular the commercial aspect, the protection of copyright and privacy.
But one of the main fears is the increase in cyberattacks. “With AI technologies such as ChatGPT readily available to the public, it is possible for hackers to generate malicious code and emails at a faster and more automated rate”warns the cybersecurity company Check Point Research.