“AI could strengthen the digital giants, but also weaken them”

Iartificial intelligence (AI) will reshuffle the digital cards or establish the power of the major players in the sector? The success of software capable of creating, from a simple written command, texts – like ChatGPT – or images – like Midjourney – initially seemed to confirm the second hypothesis. If it were possible, these systems would be even more dominated by Big Tech, such as Google, Microsoft, Meta (Facebook) or Amazon, than online research, social networks, software or e-commerce…

But in recent months, with the appearance of powerful AI models released in open source, therefore accessible to all, an alternative thesis has developed: AI could weaken the digital giants.

To support the idea of ​​a consolidation of the power of the dominant players, there is no shortage of elements: OpenAI, the start-up which created ChatGPT, has since 2019 been firmly anchored, through a partnership, to Microsoft, which would be ready to invest 10 billion dollars (9.4 billion euros). Google has bet 300 million dollars to take a stake in the start-up Anthropic. And, through their cloud hosting and services subsidiaries, Google has partnered with Cohere or C3 AI and Amazon with Stability AI. The digital giants bring computing capacity here to train gigantic models (540 billion parameters for PaLM, from Google). This is a crucial and expensive resource, as the latest chips are worth up to $40,000 apiece.

Read also: Article reserved for our subscribers Artificial intelligence would be as dangerous as ‘pandemics or nuclear war’, say industry leaders

This stranglehold is already arousing criticism, in particular from Elon Musk, for whom creating OpenAI was a way of not leaving AI in the hands of Google. “Big Tech’s dominance over the digital economy will be sealed if regulators don’t step in,” warned Sarah Myers West of the NGO AI Now Institute, in the FinancialTimes. The FTC, the US antitrust authority, said it was “vigilant”because “AI risks further consolidating the domination of the major digital players”.

Future “hybrid”

Some fear that small businesses and public research will depend on a handful of large AI models, as they depend on large platforms, social networks, mobile environments… Added to this is economic uncertainty – these models are today charged fractions of a penny per query today, but tomorrow? – as well as questions of sovereignty…

Read also: Article reserved for our subscribers Sam Altman, CEO of OpenAI: “You have to find the right balance between regulation and innovation”

But for others, the colossi of tech would rest on feet of clay: “It’s an inconvenient truth, but we at Google are not in a position to win the AI ​​sprint race. And neither is OpenAI… Open source is overtaking us, launched a Google engineer in an internal memo that leaked in early May on Semi-Analysis. Our lead is melting at an incredible speed. Open source models are faster, more adaptable by the customer and more efficient”, says the employee, pointing out that “no one will pay if a free alternative exists”. From now on, software, much smaller but powerful on specific tasks, can be trained on simple computers, even a smartphone, for a modest cost.

You have 22.99% of this article left to read. The following is for subscribers only.

source site-30