Economic effects of artificial intelligence 1/3

Learn to live with and use them

The justice of intelligence is wisdom. The wise man is not he who knows many things, but he who sees their just measure
Plato

The effects of artificial intelligences on our lives

According to a study conducted by Allied Market Research, the global generative AI market is expected to reach $10.26 billion as early as 2026, growing at a CAGR of 22% between 2019 and 2026. (Allied Market Research). But beyond just the AI sector, the entire global economy could be affected. Mc Kinsey projects $13,000 billion, global GDP growth by 2030

The applications of generative A.I. touch 3 main areas:

language production and language tools (intellectual technologies),
creativity including creation and
computer aided engineering (design), forecasting of complex phenomena and trend detection (foresight).

Creativity, engineering and foresight are shifting the computing machines to more complex analysis and projection functions, than the simple computation of data (databases, calculations, classification, clustering). These functions were previously reserved for humans.

Production of language and language tools

At the heart of intellectual technologies, one of the specificities of generative AI is the production of texts, especially computer programs. According to a study by OpenAI, large-scale language models (such as GPT-3 or GPT4) have the potential to significantly reduce software development costs by automating certain programming tasks.

For example, OpenAI used GPT-3 to generate HTML code for a basic website. This acceleration in production capabilities has not escaped the attention of businesses. A survey conducted by MIT Technology Review found that more than 60 percent of companies surveyed have already implemented or are considering using generative AI to improve their content creation process, such as ads, blog posts, and videos. (Source: MIT Technology Review).

The democratization of programming capabilities being promoted is part of a dynamic of "no-code development. The flexibility of these new tools opens the possibility for business teams, not initiated to IT development languages, to create applications thanks to visual parameterization interfaces, in a few clicks or by simple drag and drop, without any line of code. Generative AI thus opens access to software development by democratizing computer programming.

In theory, it's enough to use natural language, in practice, you need some coding skills to drive the creation process. See Using AI to its full potential can be learned!

Speeding up task completion

To increase the computational capabilities of computers, researchers at the University of California at Berkeley used generative AI. They managed to create an improved version of the most advanced AI chip to date, the Google Tensor Processing Unit (TPU). Their model, named "NASBOT," exceeded the TPU's performance in terms of speed and energy efficiency.

Players like Microsoft are investing heavily in AI accelerators to boost artificial neural networks and develop their autonomous solutions to learn ever faster. These technical gems have real-world applications, for example Photoroom, a French startup specializing in image processing, is now using an AI like midjourney to offer merchants automatic clipping, shadow display and realistic background generation. That's millions of images that can be processed simultaneously and in reduced time.

Creativity, Computer Aided Design and Engineering

As the art world is often at the forefront of new uses, contemporary artists are using generative AIs to combine already existing image backgrounds to create unique and original works of art, such as the series of works "Portrait of Edmond de Belamy" sold for $432,500 at Christie's in 2018. Now everyone seems to be able to produce creative, original and unique images without any artistic talent other than describing what they want the result to be.

It is as if knowing how to say, verbal description of the expected work, takes precedence over knowing how to do, concrete realization. At the same time this creativity is used to fabricate false images and deceive opinion. This poses various challenges of regulation.

Generative AI can still help create more realistic and immersive video games by generating worlds, characters and objects, autonomously. According to predictions, the online gaming industry could become one of the most lucrative by 2026. The market could be worth more than $300 billion, and investment is steadily increasing, even though the metaverse doesn't seem to be taking off yet.

Creativity in business is called engineering, computer-aided design. According to a study by McKinsey & Company, generative AI has the potential to generate cost savings of up to 90 percent in certain areas, including content creation, computer-aided design and supply chain management.

A variety of business uses have been catalogued, with no business function left out as detailed in this study from McKinsey. Scouting, the generation of new ideas, is particularly identified. AI's ability to produce syntheses and summaries is accessible to the widest possible audience, with the risk that online work will exteriorize confidential company data. These word processors increase employee productivity. In a survey of a call center, two groups of employees working with and without AI were observed. Those benefiting from AI showed 18% more productivity. Which reminds us, by the way, of the antiphon of value sharing when a new technology comes along: who will benefit from the productivity gain?

Hyper-rapid prototyping is now within sight. Generative AIs can help create product prototypes faster and at lower cost. For example, one shoe company used generative AI to create over 300,000 different shoe designs in just a few hours. For Tim Cook the head of Apple "AI design AI will allow creatives to focus on their real missions." Many examples of prototyping are already affecting the automotive or aerospace industry. New professions are emerging around the creation of quality instructions.

Modeling of complex phenomena

A variety of applications are emerging to address complex problems. For example, researchers are successfully predicting heat waves up to a month in advance. A major scientific breakthrough as heat waves rage.

Medical diagnoses are also difficult and sometimes generate human error. According to a study conducted by PwC, the use of generative AI provide more detailed analysis in, for example, cancer spotting. However, will we agree to trust a machine without empathy over a doctor?

Trend and market analysis

Large fashion companies such as H&M or Zara are using AI to manage their distribution chain . They also rely on AI to analyze trends and predict what styles will be popular in the future. With the increasing availability of AI tools, it is possible to analyze large amounts of data and make more informed decisions across all business registers, from design to marketing.

Also, there is no longer a need for models for long photo shoots-AI positions the character in the right setting and provides the best scenes based on target markets. What if AIs made better decisions than humans?

The Indian association Chhatra Sansad has decided to jump the gun and appoint ChatGPT as the association's president. Here's what the association's current president and CEO, Kunal Sharma, says as he steps aside for AI. " ChatGPT will use its language analytics skills to analyze market trends, identify new opportunities, and develop strategies to improve the position and development of young people in India."

Faced with this onslaught, creatives, knowledge workers, and experts with complex heuristics are being called upon. All language professions are affected, including where we did not expect it, this is also what a rabbi confided to me who notes the quality of a ChatGPT preaching in place of the usual preaching.

Illustration - Retro-futuristic couple - Midjourney

Sources

Ottho The No-Code Development https://www.ottho.co/blog/the-no-code#:~:text=Definition%20of%20No-Code%20The%20platforms%20of%20development%20No-Code,and%20to%20launch%20a%20mobile%20or%20Web%20application

Open AI https://openai.com/research

HBR France Why Generative Artificial Intelligence is a Revolution of Scale https://www.hbrfrance.fr/innovation/pourquoi-lintelligence-artificielle-generative-est-une-revolution-denvergure-60008

Wikipedia Artificial neural network Https://en.wikipedia.org/wiki/Puce_d%27acceleration_of_neuron_networks

Informatics news Microsoft is reportedly developing its own AI gas pedals https://www.informatiquenews.co.uk/microsoft-would-develop-its-own-accelerators-ia-93342

Photoroom https://www.photoroom.com/fr

BFMTV A painting made by an artificial intelligence sold for nearly 400,000 euros at auction https://www.bfmtv.com/tech/culture-web/realise-par-une-intelligence-artificielle-un-tableau-vendu-pres-de-400-000-euros-aux-encheres_AN-201810260052.html

Finance Artificial intelligence able to predict the weather https://fr.finance.yahoo.com/actualites/l-intelligence-artificielle-capable-prédire-080236298.html?

20 minutes. ChatGPT officially becomes PDF of a company in India https://www.20minutes.fr/high-tech/4023723-20230214-chatgpt-devient-officiellement-pdg-entreprise-inde

PWC How do you deploy AI at this scale? https://www.pwc.fr/fr/publications/data/comment-deployer-ia-a-l-echelle.html