Artificial intelligence is now widespread in a multitude of economic sectors. It is an integral part of business and consumer life. Technological advances and the power of AI seem limitless. Today, IAs can write articles, books, synthesize images, create works of art, identify and simulate human emotions to create avatars with empathy and emotions.
Dall-E and Mid-Journey have been making the buzz for several months because these artificial intelligences make it possible to create an image from a sentence (AI text to image). These AI parse the words used in the user’s query and then manipulate visual data to create unique images. Dall-E is a tool of the company OpenAI and Midjourney was launched by the independent research laboratory of the same name.
Midjourney also allowed Jason Allen to win an art contest in Colorado last September. Note that to access the beta version of Midjourney just go to their Discord.
This trend is also creating new jobs as a query creator (prompts) for Text to Image artificial intelligence.
For its part, PhotoRoom can extract an object from a photo taken by a smartphone to be used in an advertising poster. Runwayml allows you to modify videos from text queries, will we go towards the end of photoshop? Amazon DeepComposer allows you to create musical melodies. OpenAI with the Jukebox neural network also allows to create pieces of music and to create synthetic voices.
The power of algorithms also allows the development of no-code solutions for building applications. Tech giants have all taken this path like Microsoft Power Apps.
Suffice it to say that communications departments see these tools as an opportunity to create visuals in a very simplified manner. The synthetic media are taking power!
Through these examples we can legitimately wonder what will become human creativity if everything is delegated to artificial neural networks. In South Africa, Dabus artificial intelligence was even credited with a first patent for an AI.
This generated much debate about whether an AI and not a human being could own a patent. While this is the case in Australia or South Africa, the American courts have stipulated very clearly that patents must be held by humans. This decision is quite logical in terms of responsibility management, difficult to legally condemn an AI in case of problem. Especially since artificial intelligences are always dependent on the data they analyze and learn from. Data that still remain too often biased as demonstrated by the University of Standford and Mc Master.
Nevertheless, through these examples, we identify a computer revolution that is emerging with the multiplication of artificial intelligences of assistance whose objective will be to simplify the tasks of employees.
But in the long run, won’t employees become assistants to the neural networks themselves?