The Japan Times - New world for users and brands as ads hit AI chatbots

ChatGPT developer OpenAI began showing ads in chatbot conversations for free and low-cost users to start balancing its hundreds of billions in spending commitments with new revenue sources.

It swiftly came in for mockery from rival Anthropic, which has staked its reputation on safety and data security.

Anthropic's advertisement broadcast during last week's Super Bowl showed a man asking advice from a conversational AI, which then shoehorns advertising copy for a dating site into its otherwise relevant response.

OpenAI boss Sam Altman shot back that the clip was "clearly dishonest".

Beyond OpenAI, Microsoft has been running contextual ads and sponsored content in its Copilot AI assistant since 2023.

AI search engine Perplexity has been testing ads in the United States since 2024, while Google is also testing ads in the AI "overviews" its namesake search engine has been offering since last year.

- Data privacy -

Google has repeatedly denied wanting to run ads in its Gemini chatbot, with Demis Hassabis -- head of the search giant's DeepMind AI arm -- saying that ads "have to be handled very carefully".

"The most important thing" in AI is "trust in security and privacy, because you want to share potentially your life with that assistant," he added.

OpenAI has sought to reassure users that ChatGPT's responses will not be modified by the ads, which are shown alongside conversations rather than being integrated into them.

It has also promised not to sell user data to advertisers.

AI companies are "concerned that selling ads will scare away users," said Nate Elliott, an analyst with US data firm Emarketer.

But "when it's free, you're the product. It's a risk we're all more or less aware of already," said Jerome Malzac of AI consultancy Micropole.

"We accept it because we find value in it."

If that proves true, advertisers will be delighted to surf the AI wave as it crashes over the world's internet users.

- Game changer -

"It's going to be a game changer for the entire industry," said Justin Seibert, head of Direct Online Marketing.

"We're already seeing how high the conversion rates (interactions resulting in a purchase) are for people that are coming in from ChatGPT and the other LLMs (large language models)," he added.

AI assistants could account for up to two percent of the online advertising market by 2030, HSBC bank analysts suggested in a report.

Many brands are already prioritising visibility on the new channel, including US supermarket chain Target and software maker Adobe.

Beyond buying a spot on users' screens, companies are also pushing for their products to appear in chatbots' organic responses.

The practice is known as GEO (Generative Engine Optimisation) -- an evolution of the Search Engine Optimisation strategy during the era of Google's dominance over the web.

"We identified 90 rules that can make sure the content you create is valued by AI and spread to the right places," said Joan Burkovic, head of French GEO startup GetMint.

The company already claims 100 clients, including fashion brand Lacoste.

Malzac highlighted techniques like including references to scientific papers, adding a "frequently asked questions" section to your website, and posting information that's structured and regularly updated, Malzac said.

"If your brand isn't referenced (by chatbots) it no longer exists" for some users, he warned.

S.Ogawa--JT