In the United States, AI companies must now report their incidents to the authorities. That is good news. But one question remains for you and me: if an AI causes us harm, who do we tell?
The answer depends on what happened to you. Here are the right people to contact in France, situation by situation.
You are the victim of a scam or a hack
A phishing email written by an AI, a fake website, a hacked account, a cloned voice used to extract money from you: start with the site cybermalveillance.gouv.fr. It helps you understand what happened, gives you the first steps to take and can point you towards professionals. The 17Cyber service, available online, also offers rapid help to victims.
If you have lost money, file a complaint. For certain internet scams, the official THESEE platform lets you do so directly online.
Your personal data has been misused
A leak at an AI service, information about you used without your consent, a deletion request ignored: write first to the company, ideally to its data protection officer, whose address appears in its privacy policy.
Without a satisfactory response, you can file a complaint online with the CNIL, the French data protection authority. It is free.
Illegal content is circulating
A deepfake that humiliates you or impersonates you, hateful content, an intimate image created without your consent: report it first to the platform hosting it, then on Pharos, the official portal for reporting illegal content online, at internet-signalement.gouv.fr. And if you are the direct victim, do not hesitate to file a complaint.
A consumer problem
A misleading advert, a service that does not do what it promises, a customer service assistant that sends you round in circles: the SignalConso platform, run by the authority responsible for consumer protection, lets you report it. And in the event of a dispute, do not forget the consumer mediator, free of charge, whom we mentioned in our guide to reaching a human.
An unfair decision was made by a machine
A refused loan, a closed account, an application rejected, with no human apparently having looked at your file: the European data regulation gives you the right to request human intervention, to make your point of view heard and to contest the decision. We explained this in relation to job interviews conducted by an AI. If the company refuses, the CNIL can be contacted.
Whatever the situation, the first reflex is the same. Take dated screenshots, note the exact address of the pages concerned, export or copy the conversation with the assistant, keep the emails you receive. Do not delete anything, even what makes you uncomfortable. Without a trace, a report has little chance of succeeding. With evidence, you give the authorities what they need to act.
What we take away
Faced with an AI, you often feel alone: who do you complain to about a machine? But behind every AI there is a company, and rules that apply to it.
The AI giants must learn to be held accountable. You too have the right to demand it of them.