In the context of the unregulated uses of the ChatGPT conversational robot and the dangers of AI, economist Vincent Lorphelin invites the European Commission, in an op-ed for "Le Monde", to take the Neuralink boss at his word and take up his call for cooperation with public institutions.
Tesla CEO Elon Musk warns humanity of the dangers of artificial intelligence (AI). The lightning progress of ChatGPT makes him fear that "machines are flooding our information channels with propaganda and untruths" to the point of "losing control of our civilization".
This message should be taken very seriously. Elon Musk is probably one of the most legitimate people in the world to talk about artificial intelligence (AI). He is behind ChatGPT, autonomous cars and reusable rockets, all based on AI.
He also founded Neuralink, a company specializing in brain-controlled interfaces without the need for a mouse or joystick, and is launching a new competitor to ChatGPT. He's even allowed himself to tackle Mark Zuckerberg and Bill Gates for their "limited" knowledge of AI.undefined
A threat to which he contributes
His message is backed by hundreds of recognized experts. Eight Eastern European prime ministers are urging Big Tech [the likes of Google, Amazon, Tesla, Alphabet, Microsoft, Meta, Intel] to contain Russia's disinformation war, which aims to "destabilize our countries and weaken our democracies". NATO recognizes that, while propaganda is nothing new, its scope and intensity are growing at an alarming rate. Italy has decided to block access to ChatGPT.
The European Parliament should immediately summon Elon Musk to a hearing in Brussels, followed by his counterparts from other Big Tech companies. The Tesla CEO's message expresses a willingness to cooperate between the public and private sectors, and the scale of the threat justifies the urgent construction of countermeasures against misinformation.
The first response is to identify misinformation. We already know that the main cause of the danger stems from the way AI is programmed, resulting in a "black box" whose results nobody, not even the programmers, can explain. Malicious or military use of this tool is therefore beyond the control of its designers.
Quick solutions
Elon Musk needs to work with governments to find rapid solutions, such as the French approach, which relies on "trusted", i.e. explainable, AI. Or that of the head of diplomacy of the European Union (EU), which is based on a decentralized alert platform to react faster to infoxes.
The lack of regulation of cryptocurrencies did not stop the US Congress from putting a stop to Facebook's private currency project, libra, renamed diem. The public broadcast of the hearings took the citizen-consumer by storm, forcing Mark Zuckerberg to pledge not to launch libra anywhere in the world without US authorization. The same lever should enable the European Parliament to put the brakes on the spread of infomercials by Twitter and other social networks.
The third response anticipates proliferation beyond social networks. Elon Musk recommends the search for systems that are "more accurate, more secure, interpretable, transparent, robust, trustworthy and fair". When you're building cars, it's a good idea to equip them with controls and a dashboard to keep them under control.
Content traceability
However, this does not prevent them from being used maliciously. ChatGPTs and other generative artificial intelligences will massively build fake administrative, corporate and press sites, and intoxicate each other with content whose veracity will itself be estimated by dubious AIs.
Elon Musk adds that these systems will have to be "monitored, regulated, audited, certified", which is also required of any vehicle to drive on the road. Fortunately, he does not forget the manufacturer's and operator's responsibility for "governance" and "damage caused". As with the food industry, this heavy but necessary responsibility implies content traceability, equivalent to food traceability.
For years, France and Europe have been demanding recognition for authors, whose content has for too long been harvested for the sole benefit and control of Big Tech. These authors have finally obtained a right to remuneration and a right to object to the exploitation of their content by AIs, but these rights are struggling to be applied in practice. The hearing is intended to stabilize the liability framework for the major manufacturers and operators of generative AI, in exchange for content traceability and substantial remuneration for authors.undefined
Elon Musk's invitation is a unique opportunity to assert Europe as a regulatory power that protects democracy and copyright.