Challenges and best practices of generative AI in business

Jérôme Béranger, CEO of GoodAlgo, publishes an op-ed in Harvard Business Review France on the challenges and best practices of generative AI in business.

"Generative AI is not an end in itself: it is a lever whose risks must be mastered in order to derive lasting value from it."

Amid the enthusiasm surrounding generative AI tools (ChatGPT, Gemini, Claude…), companies must adopt a structured approach to maximise the benefits while controlling the risks. The main challenges identified:

• Confidentiality risks — data entered into these tools may feed back into training models
• Hallucinations and reliability — models sometimes generate incorrect content with apparent certainty
• Intellectual property — what about generated content and its legal reuse?
• Accountability — who is responsible for errors produced by an AI system in a business?

GoodAlgo supports companies in developing their generative AI usage policy: usage charter, team training, compliance audit and implementation of appropriate governance.