Products
Ready to get started?
No matter where you are on your CMS journey, we're here to help. Want more info or to see Glide Publishing Platform in action? We got you.
Book a demo
The best AI for publishers and media teams just got better with increased safeguards and assurances over content and usage.
Glide Publishing Platform is pleased to announce the integration of AWS Bedrock Guardrails in its industry-leading headless content management system for publishers and media, Glide CMS, giving customers increased control and capability to meet the growing demands of safety and responsibility in the use of AI.
The powerful and configurable Guardrails service is now a standard capability in Glide CMS's suite of native AI tools, Glide AI Assistant (GAIA), which brings the benefits of AI directly to publishers and media teams.
Glide GAIA already deploys as standard with industry-leading assurances over AI usage and AI content when compared to general AI systems or off-the-shelf solutions.
In addition to giving unrivalled choice over LLM usage and retention of human-in-the-loop workflows, GAIA ensures a secure ring around proprietary IP and data and prevents consumption by AI models - a critical need from creative industries and publishing businesses.
AWS Bedrock Guardrails takes that peace of mind over the use of AI to a new level.
As a top-tier AWS technology partner and participant in the AWS Voice of the Pioneer programme, Glide is invited to preview and test new services and product offerings ahead of general availability or early in the release cycle, as has been the case with AWS Bedrock Guardrails.
Combining the default protective stance of Glide GAIA with the incredible configurability of AWS Bedrock Guardrails allows customers to define powerful rules over how AI is used, with new controls over prompts and content including:
On the integration of AWS Bedrock Guardrails, Nedim Dedic, CTO of Glide Publishing Platform, said: “From the inception of Glide GAIA we have prioritized media and publishing industry concerns over IP, content control, and the retention of people in the process.
“Publishing and media companies are extremely concerned over the safe and responsible use of AI, and AWS Bedrock Guardrails helps us stay ahead of those needs. Customers can craft detailed and effective AI rulebooks specific to their organisation, giving peace of mind to them and their audiences.”
Richard Fairbairn, CPO of Glide Publishing Platform, said: “There's a reason so many media and publishing teams are cautious about the use of AI or the idea of slapping off-the-shelf AI tools on a complex CMS. A single AI headline or image can have disastrous consequences if it wipes away trust which took decades to build up or causes legal blowback. There are lots of publishers out there writing stories about how AI has caused major issues elsewhere while thinking, 'For the grace of God there go I...'.
"GAIA was built to meet these real-world priorities while helping rocket progress across the board in the other things that matter - audience and product development, better engagement, revenue growth, reduced costs, faster workflows - all while giving them critical protections over reputation, legal exposure, and their own IP."
With the global and regulatory backdrop around the use of AI in content moving so fast, the creation of AI rulebooks been pushed to the top of the agenda for media firms serious in integrating such tools.
“GAIA with Guardrails is a match made in heaven. It works just like anything else in GAIA - customers control the things they want and can change parameters and definitions quickly, while keeping cost control tight. No big set-up - you just turn it on and use it. It’s another layer of assurance for publishers to control what goes into their systems or gets put in front of their audiences," Fairbairn explained.
“Aside from all the obvious uses, it lets you do things like curate user-submitted content, set rules against providing advice or content on topics like finance or law, exclude people or topics you don’t want to cover by AI, and fact check things instantly. It’s another game-changer that puts AI power in the hands of the users, not the other way round.”
Granular content filtering of words/images/files against threat, hate, profanity etc.
A publisher may want to deploy a chatbot to a consumer-facing product which gives summaries and insight on legal matters by uploading court documents en masse into an AI system. Despite the full text and articles of evidence being fully reproduceable from a legal standpoint, the publication may decide that verbatim publishing of testimony or evidence might lead to chatbot answers which can too easily be taken out of context or be offensive, and set rules which guard against doing so.
Topic and subject filtering based on wide or fine-targeted parameters and criteria
An organisation working in the charitable or non-profit sector may want to avoid giving advice or answers on certain topics. This also applies to publishers and media in topics like financial reporting and analysis who wish to make sure they do not accidentally give financial advice, or inadvertently taking sides in a cultural or political debate outside of direct editorial guidelines.
Personal or Sensitive Data filtering, inc. PII, email, address, ID etc.
Site owners may wish to ensure that users or audiences are not able to submit any GDPR-relevant data into prompts, such as personal emails, addresses, dates of birth etc. Even where a chatbot can be tailored not to ask for such information, a user might accidentally upload personal information the company does not want in its systems - PII filtering ensures it will not be accepted and can automate or remove additional phases of analysis and filtering from workflows.
Guards against prompt manipulation and attack by malicious actors or other services
One reason publishers have been cautious about deploying products which send AI-generated content direct to audiences without human oversight is the idea that inadvertent or deliberate prompt manipulation could effectively give a third-party control over the publication’s output. Having specific safeguards in place which mitigate against this covers this off and in fact likely will become a default procurement requirement in this sector and others.
Protection against model hallucination in generated responses
Hallucinations are still one of the biggest blockers to reliance by publishers on AI systems to generate content, but being able to use Guardrails to perform contextual grounding checks gives better assurance that generated content model responses are accurate and relevant. It's a critical step before deploying chatbots, RAG outputs, and other audience-facing AI features.
Glide GAIA is a set of tools within GPP’s Glide CMS which accelerate publishing speeds and workflow efficiencies. It aids editorial teams in creating and publishing content, speeds the conversion of content for different devices and audiences, allows users to create imagery and audio files, and accelerates multilingual publishing.
In contrast to generic LLM-based AI tools, GAIA is uniquely architected to address the particular requirements of publishing and media businesses. It does so by giving them peace of mind over their content and IP, and assures humans are always in the loop before any content is published.
Glide GAIA is built on top of AWS Bedrock and other AWS systems.
To learn more about GAIA, contact pr @ gpp.io.
About Glide Publishing Platform
Glide Publishing Platform is an AI-enhanced cloud-native and API-driven headless publishing platform which removes the need for publishers to spend time and budget publishing technology to instead focus on the aspects of their business which generate revenue and response.
As a Software as a Service (SaaS), Glide provide scalability, robustness, and availability, featuring the backup enterprises' need such as self-healing infrastructure, alerts, monitoring, and round-the-clock support.
The Glide platform features a vast array of tools as standard including Live Reporting, Digital Asset Management, and automatic SEO which make the job of publishing companies and editorial teams easier, improving content output, speeding content, and significantly increasing the pace of new launched and new product development.
Glide CMS is available direct and via the AWS Marketplace.
How does Glide Publishing Platform work for you?
No matter where you are on your CMS journey, we're here to help. Want more info or to see Glide Publishing Platform in action? We got you.
Book a demo
