ITP Techblog

If you thought the tech world may take a collective pause to consider the implications of ChatGPT, think again.

This week alone we saw Microsoft demonstrate how the technology underpinning the generative AI tool from Open AI is being built into the Bing search engine and the Edge web browser.

A day later, Google unveiled Bard, its own AI-powered chatbot based on Google’s LaMDA large language model. To illustrate how high the stakes are in this face-off over AI-powered applications such as internet search, Google’s share price dropped to the tune of US$100 billion in value directly after Bard was showcased.

Microsoft CEO Satya Nadella launching the new AI-powered Bing

That’s because a promo video for Bard showed the intelligent chatbot making an embarrassing factual mistake.

Bard was asked: "What new discoveries from the James Webb Space Telescope (JWST) can I tell my 9-year old about?"

Bard offers up several answers, including the factoid that the JWST was used to take the first pictures of exoplanets, planets outside the Earth’s solar system. But that’s not true. The European Southern Observatory’s Very Large Telescope in 2004 took the first images of exoplanets, which NASA confirmed.

Garbage in, garbage out

If Bard could get a basic fact like that wrong, what else is it likely to flub, and what does it say about the integrity of Google search results. After all, as Google claims, “Bard seeks to combine the breadth of the world’s knowledge with the power, intelligence and creativity of our large language models. It draws on information from the web to provide fresh, high-quality responses”.

But the old rule applies - garbage in, garbage out. While Google’s share price largely recovered through the day of trading, investors are clearly spooked by the company’s sluggish progress in getting AI products to general release as OpenAI, with Microsoft cash flowing into it, captures the world's attention. Bard is still weeks away from debuting as an add-on to Google Search, while ChatGPT is already available on limited release as part of Bing.

Google Bard in action

Big Tech companies have been competing vigorously in AI development for years, but the debut of ChatGPT finally put the technology in the public spotlight. Finally it is tangible, its implications obvious to people. It will change how we work, how create text, videos and images, and how we search the internet.

Google claims 93% of the internet search market, a vastly profitable enterprise for the company. Now it has a motivated challenger in Microsoft, which hsa made little progress against Google in the search market to date, but which sees ChatGPT as a way to win market share of its rival. There’s a lot at stake and this AI race will only accelerate now, which is equal parts exciting and alarming.

As Vox’s Kelsey Piper wrote this week: “The usual Silicon Valley spirit — try things, see how they fail, try again — has brought us some incredibly cool consumer technology and fun websites. But it’s not how you’d build a skyscraper or a manned rocket, and powerful AI systems are much more in the latter category, where you want robust engineering for reliability”.

ChatGPT makes factual errors too. A lot of work lies ahead to improve the accuracy of these chatbots and to avoid them inadvertently spreading misinformation. As Piper also notes, the AI arms race isn’t just between Silicon Valley giants, it is operating at a national level too. China has its own large language models and its tech companies are hurrying to deploy them too.

Accelerated hype cycle

We are at the peak of heightened expectations on Gartner’s ‘hype cycle’ when it comes to generative AI, about to enter the trough of disillusionment as we realise the limitations of these systems. Big Tech, worried about a regulatory clampdown and burnt by high-profile examples of biased AI systems causing havoc in the justice system and insurance industry, has in recent years taken a very cautious approach to the release mass market AI applications.

ChatGPT has changed everything. The genie is out of the bottle at a time when we still don’t have any official regulation or oversight of how AI is used. We are accelerating through that hype cycle as everyone realises the power of this technology. We haven't seen anything like this probably since the launch of the iPhone.

But the commercial imperative to be first, to stake a claim in a burgeoning new industry comes with massive risk and in the world of AI, the consequences of getting it wrong could be devastating.