Analysis | Artificial? Yes. Intelligent? Maybe: The Great AI Chatbot Race

Remark

Here’s something you don’t see every day: Microsoft Corp. offers a hip web search function. And Google, whose search page hasn’t changed much in 24 years, is also racing to launch an equally cool revamped tool in the coming weeks. It looks like the new chat engine wars are officially on, with Microsoft announcing the long-awaited integration of OpenAI’s ChatGPT bot into Bing on Tuesday, calling it a “copilot for the web.” Google published a blog post hours earlier about its own search chatbot, called Bard. For Google in particular, it could be the riskiest strategic move it’s made in years, a metaphorical leap off the couch the company has been relaxing on for far too long.

This fight between two typically slow-moving tech giants — whose endgame represents nothing short of owning the next era of online search — will be messy and full of risk. Both companies use AI systems that are trained on billions of words on the public internet, but can also provide incorrect and even biased information. Google also risks drawing a backlash from the web publishers critical to its business.

Update: Google is already having problems with chatbot accuracy after an example used in one of Bard’s promotional samples was found to be wrong. Shares of the company fell 8% following the news.

ChatGPT sparked a wave of admiration for its creative responses to human cues when it launched last year, but since then there has been a growing concern about its understanding of facts. We don’t have statistics on how often ChatGPT gives incorrect information because OpenAI doesn’t provide those numbers. It just says that the tool gets better through regular updates. But the errors are frequent enough – between 5% and 10% of the time I’ve used it – to make users increasingly wary of all the answers.

And despite strict filters preventing the bot from making political statements or spouting hate speech, users of the popular forum Reddit have discovered how to trick ChatGPT into rants against its creators using social engineering tricks with expletives. The tool has also inexplicably used pro-Russian rhetoric when answering questions about the killing of civilians in Ukraine.

That gives us a taste of the potential stumbling blocks ahead. At first glance, it seems that Google is bolder in its use of the new technology. In the examples Microsoft has posted of the new Bing, it seems that the chatbot’s answers are relegated to the side of the page, not front and center where the usual search results remain. The bot’s replies also include footnotes and links to source material, something that ChatGPT doesn’t do, but which makes Microsoft’s tool appear significantly more reliable.

In contrast, Google’s own sample of Bard showed a single summarized answer in the middle of the page, above the search results, and with no footnotes, meaning it would be impossible for a user to identify the source. Why? Perhaps because Google is under pressure to act quickly. The company’s AI lab DeepMind is currently working on a chatbot called Sparrow that will cite sources in its answers to questions, according to a recent Time Magazine interview with DeepMind Chief Executive Officer Demis Hassabis. But Google’s new tool doesn’t use DeepMind’s technology, at least not yet.

Neither company’s demonstration showed any warning that their tool could provide inaccurate answers, as ChatGPT does. They should. “This technology is much more mature than it was a year ago, but it is still nowhere near robust enough to be ahead of the game,” said Ori Goshen, co-founder of AI21 Labs, an Israeli start-up that sells access to a large-language model that competes with OpenAI’s GPT-3.5. “Big Tech has been forced to step up and the technology isn’t there yet.”

Companies like OpenAI have tried to make their language models more accurate by adding billions of extra parameters, settings used to predict words. But making such models much larger does not mean they will become much more accurate, and some researchers suspect that the return on accuracy decreases as the models grow. Ironing out those persistent falsehoods could become a chronic challenge for both Microsoft and Google, just as self-driving car technology has been “almost” there for more than six years.

Google also risks damaging its relationship with web publishers who rely on the company’s search page to drive clicks to their sites. The company long had the technology that could have allowed it to introduce AI-powered search responses many years ago — and even talked about it in 2011 — but it never did, because that would have been a $150 billion company. disrupt that is all about getting people to click through Google search results. Now forced to act, Google searches could divert attention from other websites, drawing anger from publishers who have already accused the company of abusing its dominance in online advertising. (Disclaimer: Bloomberg LP is a web publisher that gets traffic from Google searches.)

However, publishers cannot slow down the inevitable advance of technology. Other Big Tech firms such as Meta Platforms Inc. and Amazon.com Inc. will start creating their own plays in major languages, threatening publishers of writing, music, video and more over time. The shift to adopting such models will be similar to the shift from desktops to mobile made more than a decade ago, one that companies like Facebook have accomplished in spectacular fashion, while others – like Microsoft, through the ill-fated purchase of Nokia – fluttered. This pivot is going to be a messy one, all the more so as the two leading players storm in head first.

Analysis | Artificial? Yes. Intelligent? Maybe: The Great AI Chatbot Race

Share this:

Related