Google is updating its AI-powered chatbot Gemini’s free tier to stay up with competitors in the generative AI space, such as Anthropic and OpenAI. The platform’s performance and accessibility are the main goals of the improvements.
Google unveiled the Gemini 1.5 Flash, a lightweight multimodal model, in May. It will be accessible on the web and mobile devices starting this Thursday in 40 languages and around 230 countries. Gemini 1.5 Flash, according to Google, offers improvements in latency and quality, with apparent gains in reasoning and picture recognition.
In Google’s favour, running it on the back end might also be less expensive.
Also Read: 2024 Olympics: Google Becomes Team USA’s Official AI Sponsor
Google made a point of saying at the launch of Gemini 1.5 Flash that the model was a “distilled” and extremely effective variant of Gemini 1.5 Pro, designed for “narrow,” and “high-frequency” generative AI tasks. Given the overhead associated with running a chatbot platform like Gemini (see: OpenAI’s ChatGPT bills), Google is undoubtedly keen to seize any opportunity to save costs, especially if doing so has the fortuitous side effect of improving performance in other domains.
Google claims to be extending Gemini’s context window to 32,000 tokens, or around 24,000 words, in addition to the new base model (or 48 pages of text).
Context, also called the context window, is the set of input data (text, for example) that a model takes into account before producing output (text additions, for example). Larger context models have several benefits, including the ability to summarize and reason over longer text passages and files (at least in theory) and a decreased propensity to forget previously discussed issues in the context of a chatbot.
Also Read: Google Gemini AI Reads Private Drive Docs Without User Consent: Twitter Revelation
Previously, uploading files to Gemini for analysis was only possible with Gemini Advanced, a commercial version of Gemini that was only accessible through Google’s $20/month Google One AI Premium Plan. However, Google claims that all Gemini users will soon be able to upload files from local devices and Google Drive.
Google is testing a feature that shows links to relevant online articles beneath some Gemini-generated responses to counteract hallucinations, which are situations in which a generative AI model such as Gemini 1.5 Flash fabricates information. Language: English When a paragraph produced by Gemini ends, users in certain regions will notice a “chip” icon with a link to a website (or, if you’ve allowed Gemini to access your Gmail inbox, an email) where you can go deeper.
Also Read: Google Introduces ‘Google Vids’ AI Video Creator in Limited Beta Release
The decision was made in response to information that showed Google’s generative AI models might have severe hallucinations, such as putting harmless glue in a pizza recipe or fabricating evaluations of books that were written by real people. Earlier this year, Google introduced a “double-check” feature in Gemini that aims to draw attention to claims made by Gemini that are supported or refuted by other internet sources. However, it seems that the related material connections are an attempt to increase transparency regarding the information sources that Gemini may consult.
However, Google isn’t waiting to take over the channels.
Google is launching Gemini in the European Economic Area (EEA), the United Kingdom, and Switzerland. The functionality allows users to talk in recently added languages like French, Polish, and Spanish. Previously, Gemini was only available on a limited number of devices through Messages. By selecting Gemini as a chat partner and hitting the “Start chat” button, users can bring up Gemini in Messages.
Also Read: Google Cloud AI Gemini 1.5: Flash and Pro Versions Now Available
Additionally, Google is extending Teenagers’ access to Gemini globally and launching the Gemini mobile app in other countries.
Students can sign up using their school accounts for the company’s teen-focused Gemini experience, which was launched in June, though not in all countries. That will change in the upcoming week when Gemini is made available to teenagers in every nation and area where it is typically offered to adults.
Google says that it is implementing “additional policies and safeguards” to protect teenagers in conjunction with the deployment, though it does not provide specifics. To enable adolescent users of artificial intelligence (AI) responsibly, Google is introducing an “AI literacy guide” and a new onboarding process specifically designed for teenagers.
Also Read: Google Set to Acquire Cybersecurity Firm Wiz for $23 Billion
Whether children are using generative AI technologies appropriately or inappropriately is a hotly debated topic. Google is undoubtedly keen to avoid headlines that suggest Gemini is a producer of plagiarized essays or that it can offer kids poorly thought-out advice on personal issues, so it’s doing all in its power to stop the worst from happening.
