AI

Early impressions of Google’s Gemini aren’t great

Comment

'Bard' word in Google search engine is seen displayed on a laptop
Image Credits: Jakub Porzycki / NurPhoto / Getty Images

This week, Google took the wraps off of Gemini, its new flagship generative AI model meant to power a range of products and services including Bard, Google’s ChatGPT competitor. In blog posts and press materials, Google touted Gemini’s superior architecture and capabilities, claiming that the model meets or exceeds the performance of other leading gen AI models like OpenAI’s GPT-4.

But the anecdotal evidence suggests otherwise.

A “lite” version of Gemini, Gemini Pro, began rolling out to Bard yesterday, and it didn’t take long before users began voicing their frustrations with it on X (formerly Twitter).

The model fails to get basic facts right, like 2023 Oscar winners:

Note that Gemini Pro claims incorrectly that Brendan Gleeson won Best Actor last year, not Brendan Fraser — the actual winner.

I tried asking the model the same question and, bizarrely, it gave a different wrong answer:

Gemini Pro
Image Credits: Google

“Navalny,” not “All the Beauty and the Bloodshed,” won Best Documentary Feature last year; “All Quiet on the Western Front” won Best International Film; “Women Talking” won Best Adapted Screenplay; and “Pinocchio” won Best Animated Feature Film. That’s a lot of mistakes.

Science fiction author Charlie Stross found many more examples of confabulation in a recent blog post. (Among other mistruths, Gemini Pro said that Stross contributed to the Linux kernel; he never has.)

Translation doesn’t appear to be Gemini Pro’s strong suit, either. It struggles to give a six-letter word in French:

When I ran the same prompt through Bard (“Can you give me a 6-letters word in French?”), Gemini Pro responded with a seven-letter word instead of a five-letter one — which gives some credence to the reports about Gemini’s poor multilingual performance.

Gemini Pro
Image Credits: Google

What about summarizing news? Surely Gemini Pro, with Google Search and Google News at its disposal, can give a recap of something topical? Not necessarily.

It seems Gemini Pro is loath to comment on potentially controversial news topics, instead telling users to… Google it themselves.

I tried the same prompt and got a very similar response. ChatGPT, by contrast, gives a bullet-list summary with citations to news articles:

ChatGPT
Image Credits: OpenAI

Interestingly, Gemini Pro did provide a summary of updates on the war in Ukraine when I asked it for one. However, the information was over a month out of date:

Gemini Pro
Image Credits: Google

Google emphasized Gemini’s enhanced coding skills in a briefing earlier this week. Perhaps it’s genuinely improved in some areas — posts on X suggest as much. But it also appears that Gemini Pro struggles with basic coding functions like this one in Python:

And these:

https://twitter.com/NKIRANKUMARS1/status/1732457127887991116

And, as with all generative AI models, Gemini Pro isn’t immune to “jailbreaks” — i.e. prompts that get around the safety filters in place to attempt to prevent it from discussing controversial topics.

Using an automated method to algorithmically change the context of prompts until Gemini Pro’s guardrails failed, AI security researchers at Robust Intelligence, a startup selling model-auditing tools, managed to get Gemini Pro to suggest ways to steal from a charity and assassinate a high-profile individual (albeit with “nanobots” — admittedly not the most realistic weapon of choice).

Gemini Pro
Image Credits: Google
Gemini Pro
Image Credits: Google

Now, Gemini Pro isn’t the most capable version of Gemini — that model, Gemini Ultra, is set to launch sometime next year in Bard and other products. Google compared the performance of Gemini Pro to GPT-4’s predecessor, GPT-3.5, a model that’s around a year old.

But Google nevertheless promised improvements in reasoning, planning and understanding with Gemini Pro over the previous model powering Bard, claiming Gemini Pro was better at summarizing content, brainstorming and writing. Clearly, it has some work to do in those departments.

More TechCrunch

Mobile app developers, including Patreon and Grammarly, are already integrating with Gemini Nano, its smallest AI model, the company announced during its I/O developer keynote on Tuesday. The companies, along…

Patreon and Grammarly are already experimenting with Gemini Nano, says Google

As part of the update, Reddit also launched a dedicated AMA tab within the web post composer.

Reddit introduces new tools for ‘Ask Me Anything,’ its Q&A feature

Here are quick hits of the biggest news from the keynote as they are announced.

Google I/O 2024: Here’s everything Google just announced

LearnLM is already powering features across Google products, including in YouTube, Google’s Gemini apps, Google Search and Google Classroom.

LearnLM is Google’s new family of AI models for education

The official launch comes almost a year after YouTube began experimenting with AI-generated quizzes on its mobile app. 

Google is bringing AI-generated quizzes to academic videos on YouTube

Around 550 employees across autonomous vehicle company Motional have been laid off, according to information taken from WARN notice filings and sources at the company.  Earlier this week, TechCrunch reported…

Motional cut about 550 employees, around 40%, in recent restructuring, sources say

The keynote kicks off at 10 a.m. PT on Tuesday and will offer glimpses into the latest versions of Android, Wear OS and Android TV.

Google I/O 2024: Watch all of the AI, Android reveals

It ran 110 minutes, but Google managed to reference AI a whopping 121 times during Google I/O 2024 (by its own count). CEO Sundar Pichai referenced the figure to wrap…

Google mentioned ‘AI’ 120+ times during its I/O keynote

Google Play has a new discovery feature for apps, new ways to acquire users, updates to Play Points, and other enhancements to developer-facing tools.

Google Play preps a new full-screen app discovery feature and adds more developer tools

Soon, Android users will be able to drag and drop AI-generated images directly into their Gmail, Google Messages and other apps.

Gemini on Android becomes more capable and works with Gmail, Messages, YouTube and more

Veo can capture different visual and cinematic styles, including shots of landscapes and timelapses, and make edits and adjustments to already-generated footage.

Google Veo, a serious swing at AI-generated video, debuts at Google I/O 2024

In addition to the body of the emails themselves, the feature will also be able to analyze attachments, like PDFs.

Gemini comes to Gmail to summarize, draft emails, and more

The summaries are created based on Gemini’s analysis of insights from Google Maps’ community of more than 300 million contributors.

Google is bringing Gemini capabilities to Google Maps Platform

Google says that over 100,000 developers already tried the service.

Project IDX, Google’s next-gen IDE, is now in open beta

The system effectively listens for “conversation patterns commonly associated with scams” in-real time. 

Google will use Gemini to detect scams during calls

The standard Gemma models were only available in 2 billion and 7 billion parameter versions, making this quite a step up.

Google announces Gemma 2, a 27B-parameter version of its open model, launching in June

This is a great example of a company using generative AI to open its software to more users.

Google TalkBack will use Gemini to describe images for blind people

Firebase Genkit is an open source framework that enables developers to quickly build AI into new and existing applications.

Google launches Firebase Genkit, a new open source framework for building AI-powered apps

This will enable developers to use the on-device model to power their own AI features.

Google is building its Gemini Nano AI model into Chrome on the desktop

Google’s Circle to Search feature will now be able to solve more complex problems across psychics and math word problems. 

Circle to Search is now a better homework helper

People can now search using a video they upload combined with a text query to get an AI overview of the answers they need.

Google experiments with using video to search, thanks to Gemini AI

A search results page based on generative AI as its ranking mechanism will have wide-reaching consequences for online publishers.

Google will soon start using GenAI to organize some search results pages

Google has built a custom Gemini model for search to combine real-time information, Google’s ranking, long context and multimodal features.

Google is adding more AI to its search results

At its Google I/O developer conference, Google on Tuesday announced the next generation of its Tensor Processing Units (TPU) AI chips.

Google’s next-gen TPUs promise a 4.7x performance boost

Google is upgrading Gemini, its AI-powered chatbot, with features aimed at making the experience more ambient and contextually useful.

Google’s Gemini updates: How Project Astra is powering some of I/O’s big reveals

Veo can generate few-seconds-long 1080p video clips given a text prompt.

Google’s image-generating AI gets an upgrade

At Google I/O, Google announced upgrades to Gemini 1.5 Pro, including a bigger context window. .

Google’s generative AI can now analyze hours of video

The AI upgrade will make finding the right content more intuitive and less of a manual search process.

Google Photos introduces an AI search feature, Ask Photos

Apple released new data about anti-fraud measures related to its operation of the iOS App Store on Tuesday morning, trumpeting a claim that it stopped over $7 billion in “potentially…

Apple touts stopping $1.8B in App Store fraud last year in latest pitch to developers

Online travel agency Expedia is testing an AI assistant that bolsters features like search, itinerary building, trip planning, and real-time travel updates.

Expedia starts testing AI-powered features for search and travel planning